diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-A2A_agent_time_share_bars.pdf index 6b318102f8b782cde57e02b929abe591a2ebb074..a43abd7111544c07afd14c83a5feb8c4bbe64f9a 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:352c7a14291de19dbadf16b7d392e35b4dcd714240ddb6a4db3d985018cc4ed9 +oid sha256:b77b7a12d7fa38f87f9ad6af7c31bbfec95e52053613d69b38d58e5d548747af size 32338 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-H_A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-H_A2A_agent_time_share_bars.pdf index d23468a6c9f8febe82efcc986cf891a0fbb6c6b1..fda267b61704b1672dcf982b3b90ea469536602c 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-H_A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-H_A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:5d985035ea70a80582863c76eaea70ca4d70fd7bda986035e6af5974490ee6f4 +oid sha256:fe4aa9ead165abdf4f678d741b7bf7193ae6f1ed3108821d2e08c4f359c47acd size 32009 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-MCP_agent_time_share_bars.pdf index 9f626d2d2e17f654be48059dded5ded80dea173f..7787be3e1e14d5cb761b36fa70a4ee9bd1301c27 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/BookWriter-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:2522c682d260aff425a5d719535f97485b5e6079de983d78b04a24424377dfaa +oid sha256:e60c89cca0f20593d7de9cc2a9395923b028b1eaa66472b4b633c769d0f6da7c size 32555 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder-MCP_agent_time_share_bars.pdf index b20dde1e0fb1d7edfdd51e70e93a6449d1c59182..9010ff4f521df9b402d93864265c93bdc5e2614d 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:61e834f5c5cd179d3a3db65e145bf27e66ab9a834b9da4b9d21c60e8bb234b46 +oid sha256:9cf873e9516627ab85297ecfb3b86d67ef279e099953f1b3558612517bf8f165 size 30051 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder_agent_time_share_bars.pdf index 94e2b5601774280479d0576bb15ca88ea2ba900f..b79f77806023188e5bae4b54f35b1388f67ea9a5 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/EmailResponder_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:439840b7bfdff2a98f6a4ce13ae08d6973dfb2ec18fc41363ed8287392e70117 +oid sha256:accd7bcd762f1ac7cc950c3eff1cc3d53040531981ad79040ec805d0aad9e47b size 30551 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder-MCP_agent_time_share_bars.pdf index d9b67f37d6fe40d6818e6ec777a7401e94de315c..a5cb9c6ae9434e23a2fdf52c899ffe96e27b4808 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:0660abe9a00001b0ef6f579ffe4ae1a029dbf8b7db65bdda8e89761430d01dc0 +oid sha256:1fea14d263943028e2106769be4efb70aa8eeda94e519bf80b74412617e278bb size 29679 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder_agent_time_share_bars.pdf index f05844a4e25d56190ce88b937c066c5aaf8ebee4..9c8cf2f9d3a1c11e5589bd60d47febffb6f459e4 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/GameBuilder_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:993770b799a514482161236c16b779e080ac0ff39a7d2f524759cf6c13dbe7bc +oid sha256:c7b3a2f5d70875b85546358abd61a0fbec00bbb48d44532bed66913c63b0fd48 size 29679 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-A2A_agent_time_share_bars.pdf index 97987b9a6f3d7bd71fb113b76af3137773e2bfbe..6265ebb5bc8dbb60ac0ff23e0166128722580b7a 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fc6c54966bcd84c70ad612a2e905e78770110ca19558bdfe4820de451cd8746e +oid sha256:28de1fa3e8b4cff91ebbf33c2c95f099fe026e76ea27ba6663faa8627bdffb46 size 30279 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-H_A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-H_A2A_agent_time_share_bars.pdf index dca7ee85d92cfc8f91eb185937e31304d5aba18b..7f9483c25604c6878fed4c27efa53e627f28b778 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-H_A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-H_A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:47e5a54042e5c7bba69fa5941f14d0c2dda06b78bf063e7f80ee9935d8abcf91 +oid sha256:a9f0aa687e78b4f27f854a8dfa84bacdedff046b2d20aa58a50e54810eeb719c size 30154 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-MCP_agent_time_share_bars.pdf index b46b840fddfcb72038431f271c19aefc50ab9e93..c71cfaf4f86cb3ce3f06c79217ed09ec7756d379 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/LandingPageGenerator-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:08eb027f021d7b3ca46a8ad42e3d701bed167a9ae7daab92c56d4480f463c62c +oid sha256:8ffd626506a027b54a772386b5a98dfef6cb55690c92168048bf8a3e1f4e8a3b size 30302 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator-MCP_agent_time_share_bars.pdf index 51a1444fc91b6c9f6bc3d2f3fc482256cf726a55..b1a5b04f0bfefa48112478f2877dcc0f382435b3 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:5e75dfa6166dff7bf03ed12cbed13243bcb0aa3ac205959282036c61fc583abb +oid sha256:a0852262c6b9a9b0005839ed5e0de4d754d015d13a316875bf2d309eece47f2a size 28091 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator_agent_time_share_bars.pdf index 02e0c5f8e76eb5fca7e65382e0cd707f9ee76520..4c1108d0c42627452d94ce587ed07fe8dda6fe58 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/MarkdownValidator_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:a1db303ff3d45e90d2645daa37b65d8a873a7e02f37ce1d8db629658894206d0 +oid sha256:a142faeb677d17a6c5ed91e2f6a08457c5f7d43545505039b7028e84549652f4 size 28084 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-A2A_agent_time_share_bars.pdf index df25b1765415b6a43956170363cfb733a00c7d5d..7beac33c1e011b6c313eb9fb62dd1dbf1ff34be5 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:5e565958dd0e637eb29d6a5e90c9743effcd82627ec726f202414661e6fd3adb +oid sha256:72aa35c91eabaa310e50453e7a47f39aeda505e4d94ba67a70ae2f29f3375fec size 31915 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-H_A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-H_A2A_agent_time_share_bars.pdf index 7ad7dc70cd3ece5f515683f6850ed05cb0fec34d..c05899d6231c7fbce50fd60a740ba1ca54b823d2 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-H_A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-H_A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:93b4f7f13e4c8dfb6ad38d090a29713c3a0d2d6875a3c489b67ab3454ccf39b5 +oid sha256:d9faf99d3436af7049526f35c8d3a743009b75f8e1627a019cbda1647222dbc7 size 32407 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-MCP_agent_time_share_bars.pdf index 98d8baca27b59691cff8d537b0e1165a51c66465..94e30ff783c704bacf238334863151d5dae54e65 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/RecruitmentAssistant-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:63328555742a9451ba22990df0e629c2b6d31506983a78992fd939557b03689b +oid sha256:b2217b6c41e79c64e46dae3e8f7de4c47efba123292b5fc1c75e03b6de62418a size 32822 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-A2A_agent_time_share_bars.pdf index e713190644b9078cc61fec77b48ae7c7516c6f3f..7c20aacbff0d1ef09dc976d496c1c0ae82afbfd5 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:4f0fe403e177805ea2df2281672f3e56482a3a750b21bdf3b956d57fe7f09a53 +oid sha256:0d737364bf1eab578f324d8c336b0b4e9509932cc15a885855d520a1824df9a7 size 31002 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-H_A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-H_A2A_agent_time_share_bars.pdf index 2b4abd87ee0092d833026837f1820cdda2d533ee..c0a91a8edb0fffc420b03c7b68146c1c667f5a4e 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-H_A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-H_A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:4c3883cc2865a7156a65fb2f449f2c6c34a9182cc5265c5b2783b7fc9b08f238 +oid sha256:bfed2447d542e74ea90bc9858ac2d26f467623246e6ae91f5d9e986887e9753c size 31009 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-MCP_agent_time_share_bars.pdf index 1139457ac49136b7e28c51ebfbc37598b96d457f..f76c135d470073fb7791afd47d712fc41869fffe 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/SQLAssistant-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d0b77a0a7b058185bf70b6542fc96d847c0cccebe7ad471815bb42961008fe83 +oid sha256:ea01ffa2c985f685a5da39977dcd182081f565480d81622a4327f8ddade0563d size 31563 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-A2A_agent_time_share_bars.pdf deleted file mode 100644 index 0deb384be958ecabf1bc09f89cbecb5f39935849..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-A2A_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:058694b25e1d4e48bbb35391515f2322f67f53143b8b7e1b72e04ea140a0747e -size 31000 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-A2A_mix_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-A2A_mix_agent_time_share_bars.pdf deleted file mode 100644 index f9ada0759839422f3a019286d6f98bc9fb403b8a..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-A2A_mix_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:58666a7e04f293f2cb81c57664d3bc015994935d23bf1951f6748ae0c5f4ed29 -size 31009 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-MCP_agent_time_share_bars.pdf deleted file mode 100644 index 9763577d264a364a2193d1ee33024b292ef82855..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SQL_assistant-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:f8440abf6b288228c4d7346eeb71daad657e5d7c3c323a99c35bb459ae764174 -size 31561 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-A2A_agent_time_share_bars.pdf index c8d1aea55e8ed9a26501118f9ac018744f9c9c9d..e1da013f3b11d46afda70980b42484abd51019b0 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:eb54924acb6dc283250eeb3f54f41e062b881f64a740d99159e0f6b6008500e2 +oid sha256:78b70d8d916dc51ea7e3175682ca6b2640268669033b996e902165fb27e836fd size 30951 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-H_A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-H_A2A_agent_time_share_bars.pdf index 4ddffd68c01fb4ddec056876035b05cd7966cb95..eacb99f8c63344494e28d1bfc4851721e6adb6e8 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-H_A2A_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-H_A2A_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:2ffe7b6c1b55f260acdb9a2f5636332ba4fd09b4c746f8e6abdf5c936349736a +oid sha256:779f0894ddc1c4937f2ab616c0fbc4bb40445dcb4eaec54552637f0eb46897f7 size 30727 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-MCP_agent_time_share_bars.pdf index 39a78ff7048c10439d1bec3be81cbcab87ef2ffa..8aee73061bff09929185a45f78f5a917379d8caa 100644 --- a/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-MCP_agent_time_share_bars.pdf +++ b/data/processed/RQ2/Agent-Time-UnifiedY/SocialMediaManager-MCP_agent_time_share_bars.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c6624ce5da1dd51cf6275d7f512be75c802bd36921c2316b729e5fe5050cd558 +oid sha256:a6d719c7281359aa026c8689995ebae2b6420fbd944cf4323da0efe6f749efbd size 30877 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/email_auto_responder_flow-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/email_auto_responder_flow-MCP_agent_time_share_bars.pdf deleted file mode 100644 index 63314ccdc6ac94f88d98acfa8ad605a14e937edf..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/email_auto_responder_flow-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:360b44bdd8eb33035f3939473befc67a1c4abf7811017aff705c7efe1ae3ec12 -size 30051 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/email_auto_responder_flow_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/email_auto_responder_flow_agent_time_share_bars.pdf deleted file mode 100644 index eae959256012cbc5b466c8c4c4b08072be9d54fb..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/email_auto_responder_flow_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:8b1d890f58a0a18ac444cecf4152f978fe6af32ab6a1356fbfbaca1a488e3f16 -size 30551 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/game_builder-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/game_builder-MCP_agent_time_share_bars.pdf deleted file mode 100644 index 7fcbcfac4c1ef47b75fd5ac717aece682b789bab..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/game_builder-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:71cfb4a7f039097a9fed9a0edada6f31042ed91c6145a1d76bfd56f78df0f421 -size 29679 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/game_builder_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/game_builder_agent_time_share_bars.pdf deleted file mode 100644 index d9786c425fca96231977875bebcde06e67668611..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/game_builder_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:304a93d1dd40a8a0a5da764a09354d30a3c3b4b2bdad92bbe2416f2476bab275 -size 29679 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-A2A_agent_time_share_bars.pdf deleted file mode 100644 index 665260495d3f53351becd167874dc4cab2f46964..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-A2A_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:25039327ff6272301426c66b12bf24e9e2d75522ce8c35ceb2e3745e1eb28454 -size 31947 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-A2A_mix_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-A2A_mix_agent_time_share_bars.pdf deleted file mode 100644 index 57e6628627d8ae786f0b37d57400f5c2ce099238..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-A2A_mix_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:671dbf5b4dc790fbe84b47ae50ef13c0a9f2f39919b3ba40546be70cd244da8f -size 32435 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-MCP_agent_time_share_bars.pdf deleted file mode 100644 index 6828b6fab0a0949525c7eac0d50605263bcade3d..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/intelligent_recruitment_platform-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:ed4f1f3db48c121d4079303dac69d79a7e228861567dfc48566729a2d13ccf12 -size 32856 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-A2A_agent_time_share_bars.pdf deleted file mode 100644 index 5066ab8bf3b3db0cb0f93c35af3b5aaa5140048e..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-A2A_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:5d93d87d4656932ffcb8704405f1499576867db511787d49e65f5e6b34162add -size 30279 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-A2A_mix_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-A2A_mix_agent_time_share_bars.pdf deleted file mode 100644 index a9f74c7885d245f64879d9eb67cd4443aec181ae..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-A2A_mix_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:8c1732ed0b241bf599f6b26a845bfb54c6a81833a5fb6ac9c1adee1d11a5cf2b -size 30154 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-MCP_agent_time_share_bars.pdf deleted file mode 100644 index 8f3a23f39dee152c3209638a8c0936f7ab2ae150..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/landing_page_generator-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:a718ed9320ef69ca4583027ff09297457ec56beced8ad7198777297099aafa13 -size 30302 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/markdown_validator-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/markdown_validator-MCP_agent_time_share_bars.pdf deleted file mode 100644 index 0048f97124a3c103f6ea19d7e42600c82b2c34a9..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/markdown_validator-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:d13df0645e16299e7250e9edd2152c41d1b41f7fc31c00858df4e9d142934281 -size 28091 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/markdown_validator_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/markdown_validator_agent_time_share_bars.pdf deleted file mode 100644 index d702ba7cdea1d89f486354a70f0e4d11d23247bc..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/markdown_validator_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:aced7b7b3c841d9eee1b54c9d28b7988e5ae845cffcd1d42f4cc7d29159d508d -size 28084 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-A2A_agent_time_share_bars.pdf deleted file mode 100644 index d60f222d0e574bcd782961a9bd6ebc3346812278..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-A2A_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:1f38e5904abc3173b35cd0e3f8329be3933d265b20d7b792b272ff1bd9e7ab22 -size 30951 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-A2A_mix_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-A2A_mix_agent_time_share_bars.pdf deleted file mode 100644 index 60f908d54d4d27a055116ba1bf4979f137ab503c..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-A2A_mix_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:6025e3e9c7150495dd10c8ea5b45123a8d4ad4ea2afd91249ec0f497319a7b5c -size 30727 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-MCP_agent_time_share_bars.pdf deleted file mode 100644 index cbf66d5bcdaa85b82c20f65052854b15ee79a568..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/self_evaluation_loop_flow-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:37eef742b353cabe36c462798a83c6fa1938314bb39f225601efd9ff2c13d7a9 -size 30877 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-A2A_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-A2A_agent_time_share_bars.pdf deleted file mode 100644 index b2569946173c317ddb057c52e12af397866571ab..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-A2A_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:0220f3bf73b9884a6b40db26d98f136ef116761886fb3f6bc75424d9c216aa25 -size 32350 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-A2A_mix_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-A2A_mix_agent_time_share_bars.pdf deleted file mode 100644 index 33543d175e822fb2eb8169c9d8a2cb12b2068dbd..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-A2A_mix_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:0603195eaf04d39f5dfac875e41d19edaad20950af9a4984b9f28a74c1630679 -size 32025 diff --git a/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-MCP_agent_time_share_bars.pdf b/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-MCP_agent_time_share_bars.pdf deleted file mode 100644 index 5d8cf8ca12fb50a48df09af475d5007ac9ae41fe..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/Agent-Time-UnifiedY/write_a_book_with_flows-MCP_agent_time_share_bars.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:548682059b3e8512c7e6611c5b9f02c2c11bc080bea592d50ac5b95b7e56369d -size 32572 diff --git a/data/processed/RQ2/ECDFs/A2A_vs_A2A_mix_Detailed_Comparison.md b/data/processed/RQ2/ECDFs/A2A_vs_A2A_mix_Detailed_Comparison.md deleted file mode 100644 index de684d940e9229302363f90c392beaa7c6f723c8..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/A2A_vs_A2A_mix_Detailed_Comparison.md +++ /dev/null @@ -1,6 +0,0 @@ -# A2A vs A2A_mix: Detailed Per-Project Comparison - -This document compares A2A and A2A_mix architectures for each project, showing both per-model and overall statistics. - ---- - diff --git a/data/processed/RQ2/ECDFs/A2A_vs_H_A2A_Detailed_Comparison.md b/data/processed/RQ2/ECDFs/A2A_vs_H_A2A_Detailed_Comparison.md new file mode 100644 index 0000000000000000000000000000000000000000..540d4aeafe28ce24fdef2773f86ca8b7c0140887 --- /dev/null +++ b/data/processed/RQ2/ECDFs/A2A_vs_H_A2A_Detailed_Comparison.md @@ -0,0 +1,150 @@ +# A2A vs H_A2A: Detailed Per-Project Comparison + +This document compares A2A and H_A2A architectures for each project, showing both per-model and overall statistics. + +--- + +## Overall (All Projects) + +| Project | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | --- | +| SQLAssistant | 235.62 | 175.78 | -59.84s (-25.4%) | +| RecruitmentAssistant | 501.47 | 344.65 | -156.82s (-31.3%) | +| LandingPageGenerator | 138.60 | 129.66 | -8.94s (-6.5%) | +| SocialMediaManager | 155.30 | 115.68 | -39.62s (-25.5%) | +| BookWriter | 361.87 | 264.19 | -97.68s (-27.0%) | + +## All Projects Combined (Per-Model) + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | --- | +| GPT-5 | 69.35 | 50.18 | -19.17s (-27.6%) | +| GPT-4o-mini | 89.32 | 72.46 | -16.86s (-18.9%) | +| DeepSeek-V3-1 | 258.41 | 191.05 | -67.36s (-26.1%) | +| DeepSeek-R1 | 774.85 | 642.65 | -132.20s (-17.1%) | +| Gemini-2.5-flash | 178.36 | 104.40 | -73.96s (-41.5%) | +| Gemini-2.5-flash-nothinking | 107.27 | 61.33 | -45.94s (-42.8%) | +| Qwen3-235b | 472.44 | 319.85 | -152.58s (-32.3%) | + +--- + +## SQLAssistant + +### Project-Level Summary + +### Overall Summary + +| A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | +| 235.62 | 175.78 | -59.84s (-25.4%) | + +### Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 1130.69 | 898.25 | -232.43s (-20.6%) | +| DeepSeek-V3-1 | 161.80 | 68.15 | -93.65s (-57.9%) | +| GPT-4o-mini | 33.96 | 29.00 | -4.96s (-14.6%) | +| GPT-5 | 21.10 | 28.21 | +7.11s (+33.7%) | +| Gemini-2.5-flash | 151.07 | 83.10 | -67.97s (-45.0%) | +| Gemini-2.5-flash-nothinking | 68.41 | 54.03 | -14.38s (-21.0%) | +| Qwen3-235b | 82.32 | 69.68 | -12.64s (-15.4%) | + +--- + +## RecruitmentAssistant + +### Project-Level Summary + +### Overall Summary + +| A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | +| 501.47 | 344.65 | -156.82s (-31.3%) | + +### Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 922.16 | 774.66 | -147.50s (-16.0%) | +| DeepSeek-V3-1 | 312.32 | 245.49 | -66.83s (-21.4%) | +| GPT-4o-mini | 178.62 | 144.28 | -34.34s (-19.2%) | +| GPT-5 | 67.32 | 52.98 | -14.33s (-21.3%) | +| Gemini-2.5-flash | 357.79 | 127.63 | -230.16s (-64.3%) | +| Gemini-2.5-flash-nothinking | 195.19 | 94.40 | -100.79s (-51.6%) | +| Qwen3-235b | 1476.89 | 973.11 | -503.78s (-34.1%) | + +--- + +## LandingPageGenerator + +### Project-Level Summary + +### Overall Summary + +| A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | +| 138.60 | 129.66 | -8.94s (-6.5%) | + +### Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 302.33 | 283.59 | -18.74s (-6.2%) | +| DeepSeek-V3-1 | 215.52 | 234.43 | +18.91s (+8.8%) | +| GPT-4o-mini | 44.98 | 41.31 | -3.66s (-8.1%) | +| GPT-5 | 23.26 | 30.31 | +7.05s (+30.3%) | +| Gemini-2.5-flash | 76.95 | 73.92 | -3.03s (-3.9%) | +| Gemini-2.5-flash-nothinking | 98.14 | 47.47 | -50.67s (-51.6%) | +| Qwen3-235b | 208.99 | 196.56 | -12.44s (-6.0%) | + +--- + +## SocialMediaManager + +### Project-Level Summary + +### Overall Summary + +| A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | +| 155.30 | 115.68 | -39.62s (-25.5%) | + +### Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 557.94 | 427.00 | -130.94s (-23.5%) | +| DeepSeek-V3-1 | 158.60 | 111.56 | -47.05s (-29.7%) | +| GPT-4o-mini | 87.89 | 46.16 | -41.73s (-47.5%) | +| GPT-5 | 41.51 | 27.98 | -13.53s (-32.6%) | +| Gemini-2.5-flash | 65.45 | 65.04 | -0.42s (-0.6%) | +| Gemini-2.5-flash-nothinking | 40.70 | 29.03 | -11.67s (-28.7%) | +| Qwen3-235b | 135.01 | 102.98 | -32.03s (-23.7%) | + +--- + +## BookWriter + +### Project-Level Summary + +### Overall Summary + +| A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | +| 361.87 | 264.19 | -97.68s (-27.0%) | + +### Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 961.15 | 829.76 | -131.39s (-13.7%) | +| DeepSeek-V3-1 | 443.80 | 295.61 | -148.19s (-33.4%) | +| GPT-4o-mini | 101.15 | 101.57 | +0.42s (+0.4%) | +| GPT-5 | 193.58 | 111.41 | -82.17s (-42.4%) | +| Gemini-2.5-flash | 240.55 | 172.33 | -68.22s (-28.4%) | +| Gemini-2.5-flash-nothinking | 133.91 | 81.73 | -52.18s (-39.0%) | +| Qwen3-235b | 458.96 | 256.93 | -202.03s (-44.0%) | + +--- + diff --git a/data/processed/RQ2/ECDFs/ECDF_Model_Legend_horizontal.pdf b/data/processed/RQ2/ECDFs/ECDF_Model_Legend_horizontal.pdf index f5a74ba14f98c4ab49142feec5e1c769734c56eb..aac267744c7cc48a9e11c2c00f4948aea41ec567 100644 --- a/data/processed/RQ2/ECDFs/ECDF_Model_Legend_horizontal.pdf +++ b/data/processed/RQ2/ECDFs/ECDF_Model_Legend_horizontal.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ba9a54d22913a0f3f16719060cb4ada9b2d19ba2fee991c4e143bf1b569f1170 +oid sha256:3b70c6e36875b8117bad54bbb657a55ee3ff5dd39c65f1bc30dfb66230823983 size 13138 diff --git a/data/processed/RQ2/ECDFs/Per_Project_Model_Comparison.md b/data/processed/RQ2/ECDFs/Per_Project_Model_Comparison.md index 085b3795058cc1cd732c747c35ed3409bfae9d02..cd276344019df1b37d5aa20b89c76e29364ebfdf 100644 --- a/data/processed/RQ2/ECDFs/Per_Project_Model_Comparison.md +++ b/data/processed/RQ2/ECDFs/Per_Project_Model_Comparison.md @@ -3,7 +3,7 @@ This document compares model performance within each of the 21 projects. Each project shows: -- Model statistics (mean, median, count, total time) +- Model statistics (mean) - Relative performance compared to the fastest model --- @@ -12,15 +12,15 @@ Each project shows: ### Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| GPT-5 | 45.52 | 27.00 | 1270 | 57816.36 | -| GPT-4o-mini | 69.33 | 49.05 | 1270 | 88053.45 | -| DeepSeek-V3-1 | 191.58 | 151.97 | 1270 | 243310.25 | -| DeepSeek-R1 | 589.60 | 472.63 | 1268 | 747611.27 | -| Gemini-2.5-flash | 175.03 | 93.36 | 1270 | 222285.01 | -| Gemini-2.5-flash-nothinking | 110.57 | 56.68 | 1270 | 140427.64 | -| Qwen3-235b | 307.27 | 98.62 | 1270 | 390226.65 | +| Model | Mean (s) | +| --- | --- | +| GPT-5 | 45.52 | +| GPT-4o-mini | 69.33 | +| DeepSeek-V3-1 | 191.58 | +| DeepSeek-R1 | 589.60 | +| Gemini-2.5-flash | 175.03 | +| Gemini-2.5-flash-nothinking | 110.57 | +| Qwen3-235b | 307.27 | ### Relative Performance (vs. Fastest Model) @@ -36,7 +36,7 @@ Baseline (fastest): **GPT-5** (45.52s mean) | Gemini-2.5-flash-nothinking | 110.57 | +142.9% | | Qwen3-235b | 307.27 | +574.9% | -**Slowest model:** DeepSeek-R1 (589.60s mean, +1195.1% slower than baseline) +**Slowest model:** DeepSeek-R1 (589.60s mean, +1195.1% slower than baseline GPT-5) --- @@ -44,15 +44,15 @@ Baseline (fastest): **GPT-5** (45.52s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 961.15 | 894.17 | 60 | 57669.13 | -| DeepSeek-V3-1 | 443.80 | 425.58 | 60 | 26627.97 | -| GPT-4o-mini | 101.15 | 97.92 | 60 | 6069.04 | -| GPT-5 | 193.58 | 201.17 | 60 | 11614.51 | -| Gemini-2.5-flash | 240.55 | 238.57 | 60 | 14433.15 | -| Gemini-2.5-flash-nothinking | 133.91 | 129.30 | 60 | 8034.51 | -| Qwen3-235b | 458.96 | 403.43 | 60 | 27537.69 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 961.15 | +| DeepSeek-V3-1 | 443.80 | +| GPT-4o-mini | 101.15 | +| GPT-5 | 193.58 | +| Gemini-2.5-flash | 240.55 | +| Gemini-2.5-flash-nothinking | 133.91 | +| Qwen3-235b | 458.96 | ## Relative Performance (vs. Fastest Model) @@ -68,7 +68,7 @@ Baseline (fastest): **GPT-4o-mini** (101.15s mean) | Gemini-2.5-flash-nothinking | 133.91 | +32.4% | | Qwen3-235b | 458.96 | +353.7% | -**Slowest model:** DeepSeek-R1 (961.15s mean, +850.2% slower than baseline) +**Slowest model:** DeepSeek-R1 (961.15s mean, +850.2% slower than baseline GPT-4o-mini) --- @@ -76,15 +76,15 @@ Baseline (fastest): **GPT-4o-mini** (101.15s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 829.76 | 792.88 | 60 | 49785.81 | -| DeepSeek-V3-1 | 295.61 | 283.52 | 60 | 17736.69 | -| GPT-4o-mini | 101.57 | 97.49 | 60 | 6094.34 | -| GPT-5 | 111.41 | 99.91 | 60 | 6684.57 | -| Gemini-2.5-flash | 172.33 | 160.45 | 60 | 10339.79 | -| Gemini-2.5-flash-nothinking | 81.73 | 80.19 | 60 | 4903.86 | -| Qwen3-235b | 256.93 | 217.75 | 60 | 15416.09 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 829.76 | +| DeepSeek-V3-1 | 295.61 | +| GPT-4o-mini | 101.57 | +| GPT-5 | 111.41 | +| Gemini-2.5-flash | 172.33 | +| Gemini-2.5-flash-nothinking | 81.73 | +| Qwen3-235b | 256.93 | ## Relative Performance (vs. Fastest Model) @@ -100,7 +100,7 @@ Baseline (fastest): **Gemini-2.5-flash-nothinking** (81.73s mean) | Gemini-2.5-flash-nothinking | 81.73 | +0.0% | | Qwen3-235b | 256.93 | +214.4% | -**Slowest model:** DeepSeek-R1 (829.76s mean, +915.2% slower than baseline) +**Slowest model:** DeepSeek-R1 (829.76s mean, +915.2% slower than baseline Gemini-2.5-flash-nothinking) --- @@ -108,15 +108,15 @@ Baseline (fastest): **Gemini-2.5-flash-nothinking** (81.73s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 960.29 | 818.51 | 60 | 57617.52 | -| DeepSeek-V3-1 | 395.67 | 361.70 | 60 | 23740.13 | -| GPT-4o-mini | 112.23 | 103.68 | 60 | 6733.76 | -| GPT-5 | 140.17 | 143.72 | 60 | 8409.92 | -| Gemini-2.5-flash | 183.47 | 184.77 | 60 | 11008.22 | -| Gemini-2.5-flash-nothinking | 115.47 | 108.62 | 60 | 6928.23 | -| Qwen3-235b | 442.38 | 404.23 | 60 | 26543.00 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 960.29 | +| DeepSeek-V3-1 | 395.67 | +| GPT-4o-mini | 112.23 | +| GPT-5 | 140.17 | +| Gemini-2.5-flash | 183.47 | +| Gemini-2.5-flash-nothinking | 115.47 | +| Qwen3-235b | 442.38 | ## Relative Performance (vs. Fastest Model) @@ -132,7 +132,7 @@ Baseline (fastest): **GPT-4o-mini** (112.23s mean) | Gemini-2.5-flash-nothinking | 115.47 | +2.9% | | Qwen3-235b | 442.38 | +294.2% | -**Slowest model:** DeepSeek-R1 (960.29s mean, +755.7% slower than baseline) +**Slowest model:** DeepSeek-R1 (960.29s mean, +755.7% slower than baseline GPT-4o-mini) --- @@ -140,15 +140,15 @@ Baseline (fastest): **GPT-4o-mini** (112.23s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 233.49 | 185.78 | 98 | 22882.07 | -| DeepSeek-V3-1 | 47.23 | 43.01 | 100 | 4723.35 | -| GPT-4o-mini | 17.90 | 13.83 | 100 | 1789.63 | -| GPT-5 | 9.41 | 8.42 | 100 | 941.33 | -| Gemini-2.5-flash | 180.03 | 42.84 | 100 | 18003.41 | -| Gemini-2.5-flash-nothinking | 217.85 | 32.45 | 100 | 21785.38 | -| Qwen3-235b | 37.41 | 33.02 | 100 | 3740.70 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 233.49 | +| DeepSeek-V3-1 | 47.23 | +| GPT-4o-mini | 17.90 | +| GPT-5 | 9.41 | +| Gemini-2.5-flash | 180.03 | +| Gemini-2.5-flash-nothinking | 217.85 | +| Qwen3-235b | 37.41 | ## Relative Performance (vs. Fastest Model) @@ -164,7 +164,7 @@ Baseline (fastest): **GPT-5** (9.41s mean) | Gemini-2.5-flash-nothinking | 217.85 | +2214.3% | | Qwen3-235b | 37.41 | +297.4% | -**Slowest model:** DeepSeek-R1 (233.49s mean, +2380.4% slower than baseline) +**Slowest model:** DeepSeek-R1 (233.49s mean, +2380.4% slower than baseline GPT-5) --- @@ -172,15 +172,15 @@ Baseline (fastest): **GPT-5** (9.41s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 141.80 | 125.64 | 100 | 14179.71 | -| DeepSeek-V3-1 | 44.42 | 39.60 | 100 | 4442.01 | -| GPT-4o-mini | 18.15 | 15.84 | 100 | 1814.60 | -| GPT-5 | 8.25 | 7.88 | 100 | 825.09 | -| Gemini-2.5-flash | 115.85 | 29.18 | 100 | 11585.49 | -| Gemini-2.5-flash-nothinking | 151.00 | 21.38 | 100 | 15099.93 | -| Qwen3-235b | 38.30 | 33.32 | 100 | 3830.01 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 141.80 | +| DeepSeek-V3-1 | 44.42 | +| GPT-4o-mini | 18.15 | +| GPT-5 | 8.25 | +| Gemini-2.5-flash | 115.85 | +| Gemini-2.5-flash-nothinking | 151.00 | +| Qwen3-235b | 38.30 | ## Relative Performance (vs. Fastest Model) @@ -196,7 +196,7 @@ Baseline (fastest): **GPT-5** (8.25s mean) | Gemini-2.5-flash-nothinking | 151.00 | +1730.1% | | Qwen3-235b | 38.30 | +364.2% | -**Slowest model:** Gemini-2.5-flash-nothinking (151.00s mean, +1730.1% slower than baseline) +**Slowest model:** Gemini-2.5-flash-nothinking (151.00s mean, +1730.1% slower than baseline GPT-5) --- @@ -204,15 +204,15 @@ Baseline (fastest): **GPT-5** (8.25s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 914.94 | 930.18 | 25 | 22873.47 | -| DeepSeek-V3-1 | 518.26 | 492.54 | 25 | 12956.58 | -| GPT-4o-mini | 94.49 | 87.41 | 25 | 2362.27 | -| GPT-5 | 45.35 | 45.41 | 25 | 1133.78 | -| Gemini-2.5-flash | 163.51 | 133.48 | 25 | 4087.85 | -| Gemini-2.5-flash-nothinking | 169.19 | 133.32 | 25 | 4229.82 | -| Qwen3-235b | 317.52 | 253.19 | 25 | 7938.00 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 914.94 | +| DeepSeek-V3-1 | 518.26 | +| GPT-4o-mini | 94.49 | +| GPT-5 | 45.35 | +| Gemini-2.5-flash | 163.51 | +| Gemini-2.5-flash-nothinking | 169.19 | +| Qwen3-235b | 317.52 | ## Relative Performance (vs. Fastest Model) @@ -228,7 +228,7 @@ Baseline (fastest): **GPT-5** (45.35s mean) | Gemini-2.5-flash-nothinking | 169.19 | +273.1% | | Qwen3-235b | 317.52 | +600.1% | -**Slowest model:** DeepSeek-R1 (914.94s mean, +1917.5% slower than baseline) +**Slowest model:** DeepSeek-R1 (914.94s mean, +1917.5% slower than baseline GPT-5) --- @@ -236,15 +236,15 @@ Baseline (fastest): **GPT-5** (45.35s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 827.60 | 802.12 | 25 | 20690.04 | -| DeepSeek-V3-1 | 516.71 | 462.37 | 25 | 12917.73 | -| GPT-4o-mini | 136.40 | 92.29 | 25 | 3409.88 | -| GPT-5 | 43.16 | 43.45 | 25 | 1078.90 | -| Gemini-2.5-flash | 185.74 | 185.03 | 25 | 4643.56 | -| Gemini-2.5-flash-nothinking | 120.75 | 111.85 | 25 | 3018.77 | -| Qwen3-235b | 316.86 | 313.26 | 25 | 7921.47 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 827.60 | +| DeepSeek-V3-1 | 516.71 | +| GPT-4o-mini | 136.40 | +| GPT-5 | 43.16 | +| Gemini-2.5-flash | 185.74 | +| Gemini-2.5-flash-nothinking | 120.75 | +| Qwen3-235b | 316.86 | ## Relative Performance (vs. Fastest Model) @@ -260,7 +260,7 @@ Baseline (fastest): **GPT-5** (43.16s mean) | Gemini-2.5-flash-nothinking | 120.75 | +179.8% | | Qwen3-235b | 316.86 | +634.2% | -**Slowest model:** DeepSeek-R1 (827.60s mean, +1817.7% slower than baseline) +**Slowest model:** DeepSeek-R1 (827.60s mean, +1817.7% slower than baseline GPT-5) --- @@ -268,15 +268,15 @@ Baseline (fastest): **GPT-5** (43.16s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 302.33 | 254.54 | 60 | 18139.89 | -| DeepSeek-V3-1 | 215.52 | 175.65 | 60 | 12931.05 | -| GPT-4o-mini | 44.98 | 43.88 | 60 | 2698.52 | -| GPT-5 | 23.26 | 18.74 | 60 | 1395.76 | -| Gemini-2.5-flash | 76.95 | 70.58 | 60 | 4617.19 | -| Gemini-2.5-flash-nothinking | 98.14 | 69.14 | 60 | 5888.27 | -| Qwen3-235b | 208.99 | 128.85 | 60 | 12539.60 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 302.33 | +| DeepSeek-V3-1 | 215.52 | +| GPT-4o-mini | 44.98 | +| GPT-5 | 23.26 | +| Gemini-2.5-flash | 76.95 | +| Gemini-2.5-flash-nothinking | 98.14 | +| Qwen3-235b | 208.99 | ## Relative Performance (vs. Fastest Model) @@ -292,7 +292,7 @@ Baseline (fastest): **GPT-5** (23.26s mean) | Gemini-2.5-flash-nothinking | 98.14 | +321.9% | | Qwen3-235b | 208.99 | +798.4% | -**Slowest model:** DeepSeek-R1 (302.33s mean, +1199.6% slower than baseline) +**Slowest model:** DeepSeek-R1 (302.33s mean, +1199.6% slower than baseline GPT-5) --- @@ -300,15 +300,15 @@ Baseline (fastest): **GPT-5** (23.26s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 283.59 | 240.55 | 60 | 17015.44 | -| DeepSeek-V3-1 | 234.43 | 193.30 | 60 | 14065.93 | -| GPT-4o-mini | 41.31 | 39.94 | 60 | 2478.65 | -| GPT-5 | 30.31 | 32.22 | 60 | 1818.49 | -| Gemini-2.5-flash | 73.92 | 52.93 | 60 | 4435.47 | -| Gemini-2.5-flash-nothinking | 47.47 | 42.83 | 60 | 2848.08 | -| Qwen3-235b | 196.56 | 119.09 | 60 | 11793.38 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 283.59 | +| DeepSeek-V3-1 | 234.43 | +| GPT-4o-mini | 41.31 | +| GPT-5 | 30.31 | +| Gemini-2.5-flash | 73.92 | +| Gemini-2.5-flash-nothinking | 47.47 | +| Qwen3-235b | 196.56 | ## Relative Performance (vs. Fastest Model) @@ -324,7 +324,7 @@ Baseline (fastest): **GPT-5** (30.31s mean) | Gemini-2.5-flash-nothinking | 47.47 | +56.6% | | Qwen3-235b | 196.56 | +548.5% | -**Slowest model:** DeepSeek-R1 (283.59s mean, +835.7% slower than baseline) +**Slowest model:** DeepSeek-R1 (283.59s mean, +835.7% slower than baseline GPT-5) --- @@ -332,15 +332,15 @@ Baseline (fastest): **GPT-5** (30.31s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 274.53 | 246.57 | 60 | 16471.81 | -| DeepSeek-V3-1 | 135.84 | 119.45 | 60 | 8150.36 | -| GPT-4o-mini | 55.93 | 54.65 | 60 | 3355.75 | -| GPT-5 | 23.98 | 23.42 | 60 | 1438.98 | -| Gemini-2.5-flash | 91.42 | 87.19 | 60 | 5484.96 | -| Gemini-2.5-flash-nothinking | 97.29 | 69.47 | 60 | 5837.31 | -| Qwen3-235b | 92.07 | 86.16 | 60 | 5524.20 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 274.53 | +| DeepSeek-V3-1 | 135.84 | +| GPT-4o-mini | 55.93 | +| GPT-5 | 23.98 | +| Gemini-2.5-flash | 91.42 | +| Gemini-2.5-flash-nothinking | 97.29 | +| Qwen3-235b | 92.07 | ## Relative Performance (vs. Fastest Model) @@ -356,7 +356,7 @@ Baseline (fastest): **GPT-5** (23.98s mean) | Gemini-2.5-flash-nothinking | 97.29 | +305.7% | | Qwen3-235b | 92.07 | +283.9% | -**Slowest model:** DeepSeek-R1 (274.53s mean, +1044.7% slower than baseline) +**Slowest model:** DeepSeek-R1 (274.53s mean, +1044.7% slower than baseline GPT-5) --- @@ -364,15 +364,15 @@ Baseline (fastest): **GPT-5** (23.98s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 147.98 | 128.24 | 60 | 8878.66 | -| DeepSeek-V3-1 | 10.29 | 8.95 | 60 | 617.43 | -| GPT-4o-mini | 8.47 | 3.01 | 60 | 508.36 | -| GPT-5 | 7.32 | 2.60 | 60 | 439.11 | -| Gemini-2.5-flash | 554.35 | 221.42 | 60 | 33260.97 | -| Gemini-2.5-flash-nothinking | 132.33 | 7.61 | 60 | 7939.80 | -| Qwen3-235b | 9.54 | 7.38 | 60 | 572.29 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 147.98 | +| DeepSeek-V3-1 | 10.29 | +| GPT-4o-mini | 8.47 | +| GPT-5 | 7.32 | +| Gemini-2.5-flash | 554.35 | +| Gemini-2.5-flash-nothinking | 132.33 | +| Qwen3-235b | 9.54 | ## Relative Performance (vs. Fastest Model) @@ -388,7 +388,7 @@ Baseline (fastest): **GPT-5** (7.32s mean) | Gemini-2.5-flash-nothinking | 132.33 | +1708.2% | | Qwen3-235b | 9.54 | +30.3% | -**Slowest model:** Gemini-2.5-flash (554.35s mean, +7474.7% slower than baseline) +**Slowest model:** Gemini-2.5-flash (554.35s mean, +7474.7% slower than baseline GPT-5) --- @@ -396,15 +396,15 @@ Baseline (fastest): **GPT-5** (7.32s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 129.20 | 113.74 | 60 | 7752.24 | -| DeepSeek-V3-1 | 10.69 | 9.03 | 60 | 641.59 | -| GPT-4o-mini | 3.58 | 3.13 | 60 | 214.72 | -| GPT-5 | 2.98 | 2.80 | 60 | 179.05 | -| Gemini-2.5-flash | 457.36 | 251.42 | 60 | 27441.47 | -| Gemini-2.5-flash-nothinking | 168.14 | 11.26 | 60 | 10088.45 | -| Qwen3-235b | 10.49 | 7.17 | 60 | 629.39 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 129.20 | +| DeepSeek-V3-1 | 10.69 | +| GPT-4o-mini | 3.58 | +| GPT-5 | 2.98 | +| Gemini-2.5-flash | 457.36 | +| Gemini-2.5-flash-nothinking | 168.14 | +| Qwen3-235b | 10.49 | ## Relative Performance (vs. Fastest Model) @@ -420,7 +420,7 @@ Baseline (fastest): **GPT-5** (2.98s mean) | Gemini-2.5-flash-nothinking | 168.14 | +5534.3% | | Qwen3-235b | 10.49 | +251.5% | -**Slowest model:** Gemini-2.5-flash (457.36s mean, +15225.7% slower than baseline) +**Slowest model:** Gemini-2.5-flash (457.36s mean, +15225.7% slower than baseline GPT-5) --- @@ -428,15 +428,15 @@ Baseline (fastest): **GPT-5** (2.98s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 922.16 | 669.36 | 60 | 55329.75 | -| DeepSeek-V3-1 | 312.32 | 300.68 | 60 | 18739.34 | -| GPT-4o-mini | 178.62 | 174.95 | 60 | 10717.18 | -| GPT-5 | 67.32 | 66.63 | 60 | 4038.91 | -| Gemini-2.5-flash | 357.79 | 300.61 | 60 | 21467.37 | -| Gemini-2.5-flash-nothinking | 195.19 | 194.60 | 60 | 11711.41 | -| Qwen3-235b | 1476.89 | 1057.75 | 60 | 88613.16 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 922.16 | +| DeepSeek-V3-1 | 312.32 | +| GPT-4o-mini | 178.62 | +| GPT-5 | 67.32 | +| Gemini-2.5-flash | 357.79 | +| Gemini-2.5-flash-nothinking | 195.19 | +| Qwen3-235b | 1476.89 | ## Relative Performance (vs. Fastest Model) @@ -452,7 +452,7 @@ Baseline (fastest): **GPT-5** (67.32s mean) | Gemini-2.5-flash-nothinking | 195.19 | +190.0% | | Qwen3-235b | 1476.89 | +2094.0% | -**Slowest model:** Qwen3-235b (1476.89s mean, +2094.0% slower than baseline) +**Slowest model:** Qwen3-235b (1476.89s mean, +2094.0% slower than baseline GPT-5) --- @@ -460,15 +460,15 @@ Baseline (fastest): **GPT-5** (67.32s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 774.66 | 644.06 | 60 | 46479.85 | -| DeepSeek-V3-1 | 245.49 | 249.80 | 60 | 14729.69 | -| GPT-4o-mini | 144.28 | 131.84 | 60 | 8656.53 | -| GPT-5 | 52.98 | 50.55 | 60 | 3179.00 | -| Gemini-2.5-flash | 127.63 | 120.13 | 60 | 7657.88 | -| Gemini-2.5-flash-nothinking | 94.40 | 94.66 | 60 | 5664.18 | -| Qwen3-235b | 973.11 | 912.25 | 60 | 58386.42 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 774.66 | +| DeepSeek-V3-1 | 245.49 | +| GPT-4o-mini | 144.28 | +| GPT-5 | 52.98 | +| Gemini-2.5-flash | 127.63 | +| Gemini-2.5-flash-nothinking | 94.40 | +| Qwen3-235b | 973.11 | ## Relative Performance (vs. Fastest Model) @@ -484,7 +484,7 @@ Baseline (fastest): **GPT-5** (52.98s mean) | Gemini-2.5-flash-nothinking | 94.40 | +78.2% | | Qwen3-235b | 973.11 | +1736.6% | -**Slowest model:** Qwen3-235b (973.11s mean, +1736.6% slower than baseline) +**Slowest model:** Qwen3-235b (973.11s mean, +1736.6% slower than baseline GPT-5) --- @@ -492,15 +492,15 @@ Baseline (fastest): **GPT-5** (52.98s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 839.96 | 711.17 | 60 | 50397.41 | -| DeepSeek-V3-1 | 339.42 | 328.97 | 60 | 20365.50 | -| GPT-4o-mini | 129.31 | 107.80 | 60 | 7758.66 | -| GPT-5 | 59.50 | 57.94 | 60 | 3570.10 | -| Gemini-2.5-flash | 201.33 | 185.48 | 60 | 12079.56 | -| Gemini-2.5-flash-nothinking | 163.09 | 156.11 | 60 | 9785.44 | -| Qwen3-235b | 1381.73 | 1226.34 | 60 | 82903.53 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 839.96 | +| DeepSeek-V3-1 | 339.42 | +| GPT-4o-mini | 129.31 | +| GPT-5 | 59.50 | +| Gemini-2.5-flash | 201.33 | +| Gemini-2.5-flash-nothinking | 163.09 | +| Qwen3-235b | 1381.73 | ## Relative Performance (vs. Fastest Model) @@ -516,7 +516,7 @@ Baseline (fastest): **GPT-5** (59.50s mean) | Gemini-2.5-flash-nothinking | 163.09 | +174.1% | | Qwen3-235b | 1381.73 | +2222.2% | -**Slowest model:** Qwen3-235b (1381.73s mean, +2222.2% slower than baseline) +**Slowest model:** Qwen3-235b (1381.73s mean, +2222.2% slower than baseline GPT-5) --- @@ -524,15 +524,15 @@ Baseline (fastest): **GPT-5** (59.50s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 1130.69 | 1048.39 | 60 | 67841.22 | -| DeepSeek-V3-1 | 161.80 | 144.55 | 60 | 9707.97 | -| GPT-4o-mini | 33.96 | 30.11 | 60 | 2037.42 | -| GPT-5 | 21.10 | 21.42 | 60 | 1266.15 | -| Gemini-2.5-flash | 151.07 | 118.63 | 60 | 9063.97 | -| Gemini-2.5-flash-nothinking | 68.41 | 53.95 | 60 | 4104.37 | -| Qwen3-235b | 82.32 | 75.72 | 60 | 4939.32 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 1130.69 | +| DeepSeek-V3-1 | 161.80 | +| GPT-4o-mini | 33.96 | +| GPT-5 | 21.10 | +| Gemini-2.5-flash | 151.07 | +| Gemini-2.5-flash-nothinking | 68.41 | +| Qwen3-235b | 82.32 | ## Relative Performance (vs. Fastest Model) @@ -548,7 +548,7 @@ Baseline (fastest): **GPT-5** (21.10s mean) | Gemini-2.5-flash-nothinking | 68.41 | +224.2% | | Qwen3-235b | 82.32 | +290.1% | -**Slowest model:** DeepSeek-R1 (1130.69s mean, +5258.1% slower than baseline) +**Slowest model:** DeepSeek-R1 (1130.69s mean, +5258.1% slower than baseline GPT-5) --- @@ -556,15 +556,15 @@ Baseline (fastest): **GPT-5** (21.10s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 898.25 | 659.13 | 60 | 53895.23 | -| DeepSeek-V3-1 | 68.15 | 57.57 | 60 | 4089.18 | -| GPT-4o-mini | 29.00 | 22.22 | 60 | 1739.97 | -| GPT-5 | 28.21 | 23.74 | 60 | 1692.83 | -| Gemini-2.5-flash | 83.10 | 64.11 | 60 | 4985.79 | -| Gemini-2.5-flash-nothinking | 54.03 | 29.85 | 60 | 3241.76 | -| Qwen3-235b | 69.68 | 57.27 | 60 | 4181.03 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 898.25 | +| DeepSeek-V3-1 | 68.15 | +| GPT-4o-mini | 29.00 | +| GPT-5 | 28.21 | +| Gemini-2.5-flash | 83.10 | +| Gemini-2.5-flash-nothinking | 54.03 | +| Qwen3-235b | 69.68 | ## Relative Performance (vs. Fastest Model) @@ -580,7 +580,7 @@ Baseline (fastest): **GPT-5** (28.21s mean) | Gemini-2.5-flash-nothinking | 54.03 | +91.5% | | Qwen3-235b | 69.68 | +147.0% | -**Slowest model:** DeepSeek-R1 (898.25s mean, +3083.7% slower than baseline) +**Slowest model:** DeepSeek-R1 (898.25s mean, +3083.7% slower than baseline GPT-5) --- @@ -588,15 +588,15 @@ Baseline (fastest): **GPT-5** (28.21s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 941.69 | 908.10 | 60 | 56501.59 | -| DeepSeek-V3-1 | 112.55 | 103.47 | 60 | 6753.21 | -| GPT-4o-mini | 48.84 | 46.84 | 60 | 2930.67 | -| GPT-5 | 26.09 | 26.63 | 60 | 1565.16 | -| Gemini-2.5-flash | 104.38 | 61.00 | 60 | 6262.88 | -| Gemini-2.5-flash-nothinking | 46.95 | 29.87 | 60 | 2816.70 | -| Qwen3-235b | 89.03 | 80.44 | 60 | 5341.82 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 941.69 | +| DeepSeek-V3-1 | 112.55 | +| GPT-4o-mini | 48.84 | +| GPT-5 | 26.09 | +| Gemini-2.5-flash | 104.38 | +| Gemini-2.5-flash-nothinking | 46.95 | +| Qwen3-235b | 89.03 | ## Relative Performance (vs. Fastest Model) @@ -612,7 +612,7 @@ Baseline (fastest): **GPT-5** (26.09s mean) | Gemini-2.5-flash-nothinking | 46.95 | +80.0% | | Qwen3-235b | 89.03 | +241.3% | -**Slowest model:** DeepSeek-R1 (941.69s mean, +3509.9% slower than baseline) +**Slowest model:** DeepSeek-R1 (941.69s mean, +3509.9% slower than baseline GPT-5) --- @@ -620,15 +620,15 @@ Baseline (fastest): **GPT-5** (26.09s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 557.94 | 497.03 | 60 | 33476.23 | -| DeepSeek-V3-1 | 158.60 | 151.02 | 60 | 9516.19 | -| GPT-4o-mini | 87.89 | 75.81 | 60 | 5273.64 | -| GPT-5 | 41.51 | 33.49 | 60 | 2490.59 | -| Gemini-2.5-flash | 65.45 | 58.31 | 60 | 3927.28 | -| Gemini-2.5-flash-nothinking | 40.70 | 37.84 | 60 | 2441.80 | -| Qwen3-235b | 135.01 | 113.26 | 60 | 8100.81 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 557.94 | +| DeepSeek-V3-1 | 158.60 | +| GPT-4o-mini | 87.89 | +| GPT-5 | 41.51 | +| Gemini-2.5-flash | 65.45 | +| Gemini-2.5-flash-nothinking | 40.70 | +| Qwen3-235b | 135.01 | ## Relative Performance (vs. Fastest Model) @@ -644,7 +644,7 @@ Baseline (fastest): **Gemini-2.5-flash-nothinking** (40.70s mean) | Gemini-2.5-flash-nothinking | 40.70 | +0.0% | | Qwen3-235b | 135.01 | +231.8% | -**Slowest model:** DeepSeek-R1 (557.94s mean, +1271.0% slower than baseline) +**Slowest model:** DeepSeek-R1 (557.94s mean, +1271.0% slower than baseline Gemini-2.5-flash-nothinking) --- @@ -652,15 +652,15 @@ Baseline (fastest): **Gemini-2.5-flash-nothinking** (40.70s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 427.00 | 316.96 | 60 | 25620.06 | -| DeepSeek-V3-1 | 111.56 | 109.53 | 60 | 6693.43 | -| GPT-4o-mini | 46.16 | 38.91 | 60 | 2769.74 | -| GPT-5 | 27.98 | 25.56 | 60 | 1678.90 | -| Gemini-2.5-flash | 65.04 | 54.59 | 60 | 3902.31 | -| Gemini-2.5-flash-nothinking | 29.03 | 23.80 | 60 | 1741.76 | -| Qwen3-235b | 102.98 | 70.84 | 60 | 6178.94 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 427.00 | +| DeepSeek-V3-1 | 111.56 | +| GPT-4o-mini | 46.16 | +| GPT-5 | 27.98 | +| Gemini-2.5-flash | 65.04 | +| Gemini-2.5-flash-nothinking | 29.03 | +| Qwen3-235b | 102.98 | ## Relative Performance (vs. Fastest Model) @@ -676,7 +676,7 @@ Baseline (fastest): **GPT-5** (27.98s mean) | Gemini-2.5-flash-nothinking | 29.03 | +3.7% | | Qwen3-235b | 102.98 | +268.0% | -**Slowest model:** DeepSeek-R1 (427.00s mean, +1426.0% slower than baseline) +**Slowest model:** DeepSeek-R1 (427.00s mean, +1426.0% slower than baseline GPT-5) --- @@ -684,15 +684,15 @@ Baseline (fastest): **GPT-5** (27.98s mean) ## Model Statistics Summary -| Model | Mean (s) | Median (s) | Count | Total (s) | -| --- | --- | --- | --- | --- | -| DeepSeek-R1 | 735.24 | 628.71 | 60 | 44114.13 | -| DeepSeek-V3-1 | 219.42 | 197.26 | 60 | 13164.95 | -| GPT-4o-mini | 144.00 | 111.81 | 60 | 8640.11 | -| GPT-5 | 39.59 | 33.41 | 60 | 2375.23 | -| Gemini-2.5-flash | 59.94 | 55.27 | 60 | 3596.43 | -| Gemini-2.5-flash-nothinking | 38.63 | 37.92 | 60 | 2317.81 | -| Qwen3-235b | 126.60 | 105.77 | 60 | 7595.78 | +| Model | Mean (s) | +| --- | --- | +| DeepSeek-R1 | 735.24 | +| DeepSeek-V3-1 | 219.42 | +| GPT-4o-mini | 144.00 | +| GPT-5 | 39.59 | +| Gemini-2.5-flash | 59.94 | +| Gemini-2.5-flash-nothinking | 38.63 | +| Qwen3-235b | 126.60 | ## Relative Performance (vs. Fastest Model) @@ -708,7 +708,7 @@ Baseline (fastest): **Gemini-2.5-flash-nothinking** (38.63s mean) | Gemini-2.5-flash-nothinking | 38.63 | +0.0% | | Qwen3-235b | 126.60 | +227.7% | -**Slowest model:** DeepSeek-R1 (735.24s mean, +1803.3% slower than baseline) +**Slowest model:** DeepSeek-R1 (735.24s mean, +1803.3% slower than baseline Gemini-2.5-flash-nothinking) --- diff --git a/data/processed/RQ2/ECDFs/Time_Comparison_A2A_vs_A2A_mix.md b/data/processed/RQ2/ECDFs/Time_Comparison_A2A_vs_A2A_mix.md deleted file mode 100644 index 84cc10ac29154aeccf21d383bd44d26dad168db4..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/Time_Comparison_A2A_vs_A2A_mix.md +++ /dev/null @@ -1,38 +0,0 @@ -# Overall A2A vs A2A_mix Time Comparison - -Averaged across all projects: SQL_assistant, intelligent_recruitment_platform, landing_page_generator, self_evaluation_loop_flow, write_a_book_with_flows - -## Framework-Level Comparison (All Models Averaged) - -| A2A Mean (s) | A2A_mix Mean (s) | Diff (A2A-A2A_mix) | -| --- | --- | --- | - ---- - -## Per-Model Summary - -| Model | A2A Mean (s) | A2A_mix Mean (s) | Diff (A2A-A2A_mix) | -| --- | --- | --- | --- | - ---- - -# SQL_assistant: A2A vs A2A_mix - -_Data not available for comparison_ - -# intelligent_recruitment_platform: A2A vs A2A_mix - -_Data not available for comparison_ - -# landing_page_generator: A2A vs A2A_mix - -_Data not available for comparison_ - -# self_evaluation_loop_flow: A2A vs A2A_mix - -_Data not available for comparison_ - -# write_a_book_with_flows: A2A vs A2A_mix - -_Data not available for comparison_ - diff --git a/data/processed/RQ2/ECDFs/Time_Comparison_A2A_vs_H_A2A.md b/data/processed/RQ2/ECDFs/Time_Comparison_A2A_vs_H_A2A.md new file mode 100644 index 0000000000000000000000000000000000000000..c939d0f14233021bedaa3f0ffb80816d1304794f --- /dev/null +++ b/data/processed/RQ2/ECDFs/Time_Comparison_A2A_vs_H_A2A.md @@ -0,0 +1,136 @@ +# Overall A2A vs H_A2A Time Comparison + +Averaged across all projects: SQLAssistant, RecruitmentAssistant, LandingPageGenerator, SocialMediaManager, BookWriter + +## Framework-Level Comparison (All Models Averaged) + +| A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | +| 278.57 | 205.99 | +72.58s (+35.2%) | + +--- + +## Per-Model Summary + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 774.85 | 642.65 | +132.20s (+20.6%) | +| DeepSeek-V3-1 | 258.41 | 191.05 | +67.36s (+35.3%) | +| GPT-4o-mini | 89.32 | 72.46 | +16.86s (+23.3%) | +| GPT-5 | 69.35 | 50.18 | +19.17s (+38.2%) | +| Gemini-2.5-flash | 178.36 | 104.40 | +73.96s (+70.8%) | +| Gemini-2.5-flash-nothinking | 107.27 | 61.33 | +45.94s (+74.9%) | +| Qwen3-235b | 472.44 | 319.85 | +152.58s (+47.7%) | + +--- + +# SQLAssistant: A2A vs H_A2A + +## Overall Summary (Averaged Across All Models) + +| A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | +| 235.62 | 175.78 | +59.84s (+34.0%) | + +--- + +## Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 1130.69 | 898.25 | +232.43s (+25.9%) | +| DeepSeek-V3-1 | 161.80 | 68.15 | +93.65s (+137.4%) | +| GPT-4o-mini | 33.96 | 29.00 | +4.96s (+17.1%) | +| GPT-5 | 21.10 | 28.21 | -7.11s (-25.2%) | +| Gemini-2.5-flash | 151.07 | 83.10 | +67.97s (+81.8%) | +| Gemini-2.5-flash-nothinking | 68.41 | 54.03 | +14.38s (+26.6%) | +| Qwen3-235b | 82.32 | 69.68 | +12.64s (+18.1%) | + +# RecruitmentAssistant: A2A vs H_A2A + +## Overall Summary (Averaged Across All Models) + +| A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | +| 501.47 | 344.65 | +156.82s (+45.5%) | + +--- + +## Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 922.16 | 774.66 | +147.50s (+19.0%) | +| DeepSeek-V3-1 | 312.32 | 245.49 | +66.83s (+27.2%) | +| GPT-4o-mini | 178.62 | 144.28 | +34.34s (+23.8%) | +| GPT-5 | 67.32 | 52.98 | +14.33s (+27.0%) | +| Gemini-2.5-flash | 357.79 | 127.63 | +230.16s (+180.3%) | +| Gemini-2.5-flash-nothinking | 195.19 | 94.40 | +100.79s (+106.8%) | +| Qwen3-235b | 1476.89 | 973.11 | +503.78s (+51.8%) | + +# LandingPageGenerator: A2A vs H_A2A + +## Overall Summary (Averaged Across All Models) + +| A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | +| 138.60 | 129.66 | +8.94s (+6.9%) | + +--- + +## Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 302.33 | 283.59 | +18.74s (+6.6%) | +| DeepSeek-V3-1 | 215.52 | 234.43 | -18.91s (-8.1%) | +| GPT-4o-mini | 44.98 | 41.31 | +3.66s (+8.9%) | +| GPT-5 | 23.26 | 30.31 | -7.05s (-23.2%) | +| Gemini-2.5-flash | 76.95 | 73.92 | +3.03s (+4.1%) | +| Gemini-2.5-flash-nothinking | 98.14 | 47.47 | +50.67s (+106.7%) | +| Qwen3-235b | 208.99 | 196.56 | +12.44s (+6.3%) | + +# SocialMediaManager: A2A vs H_A2A + +## Overall Summary (Averaged Across All Models) + +| A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | +| 155.30 | 115.68 | +39.62s (+34.3%) | + +--- + +## Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 557.94 | 427.00 | +130.94s (+30.7%) | +| DeepSeek-V3-1 | 158.60 | 111.56 | +47.05s (+42.2%) | +| GPT-4o-mini | 87.89 | 46.16 | +41.73s (+90.4%) | +| GPT-5 | 41.51 | 27.98 | +13.53s (+48.3%) | +| Gemini-2.5-flash | 65.45 | 65.04 | +0.42s (+0.6%) | +| Gemini-2.5-flash-nothinking | 40.70 | 29.03 | +11.67s (+40.2%) | +| Qwen3-235b | 135.01 | 102.98 | +32.03s (+31.1%) | + +# BookWriter: A2A vs H_A2A + +## Overall Summary (Averaged Across All Models) + +| A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | +| 361.87 | 264.19 | +97.68s (+37.0%) | + +--- + +## Per-Model Comparison + +| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (A2A-H_A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 961.15 | 829.76 | +131.39s (+15.8%) | +| DeepSeek-V3-1 | 443.80 | 295.61 | +148.19s (+50.1%) | +| GPT-4o-mini | 101.15 | 101.57 | -0.42s (-0.4%) | +| GPT-5 | 193.58 | 111.41 | +82.17s (+73.8%) | +| Gemini-2.5-flash | 240.55 | 172.33 | +68.22s (+39.6%) | +| Gemini-2.5-flash-nothinking | 133.91 | 81.73 | +52.18s (+63.8%) | +| Qwen3-235b | 458.96 | 256.93 | +202.03s (+78.6%) | + diff --git a/data/processed/RQ2/ECDFs/Time_Comparison_MCP_vs_A2A.md b/data/processed/RQ2/ECDFs/Time_Comparison_MCP_vs_A2A.md index 97cc224a70d6d70b357cd2f64b34c8700bccd7bc..913f81bce04c82cee31086ad218a3a35c4c82e3c 100644 --- a/data/processed/RQ2/ECDFs/Time_Comparison_MCP_vs_A2A.md +++ b/data/processed/RQ2/ECDFs/Time_Comparison_MCP_vs_A2A.md @@ -1,11 +1,12 @@ # Overall MCP vs A2A Time Comparison -Averaged across all projects: SQL_assistant, intelligent_recruitment_platform, landing_page_generator, self_evaluation_loop_flow, write_a_book_with_flows +Averaged across all projects: SQLAssistant, RecruitmentAssistant, LandingPageGenerator, SocialMediaManager, BookWriter ## Framework-Level Comparison (All Models Averaged) | MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | | --- | --- | --- | +| 256.23 | 278.57 | -22.34s (-8.0%) | --- @@ -13,26 +14,123 @@ Averaged across all projects: SQL_assistant, intelligent_recruitment_platform, l | Model | MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | | --- | --- | --- | --- | +| DeepSeek-R1 | 750.34 | 774.85 | -24.51s (-3.2%) | +| DeepSeek-V3-1 | 240.58 | 258.41 | -17.83s (-6.9%) | +| GPT-4o-mini | 98.06 | 89.32 | +8.74s (+9.8%) | +| GPT-5 | 57.86 | 69.35 | -11.49s (-16.6%) | +| Gemini-2.5-flash | 128.11 | 178.36 | -50.26s (-28.2%) | +| Gemini-2.5-flash-nothinking | 92.28 | 107.27 | -14.98s (-14.0%) | +| Qwen3-235b | 426.36 | 472.44 | -46.07s (-9.8%) | --- -# SQL_assistant: MCP vs A2A +# SQLAssistant: MCP vs A2A -_Data not available for comparison_ +## Overall Summary (Averaged Across All Models) -# intelligent_recruitment_platform: MCP vs A2A +| MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | +| 195.65 | 235.62 | -39.97s (-17.0%) | + +--- + +## Per-Model Comparison -_Data not available for comparison_ +| Model | MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 941.69 | 1130.69 | -188.99s (-16.7%) | +| DeepSeek-V3-1 | 112.55 | 161.80 | -49.25s (-30.4%) | +| GPT-4o-mini | 48.84 | 33.96 | +14.89s (+43.8%) | +| GPT-5 | 26.09 | 21.10 | +4.98s (+23.6%) | +| Gemini-2.5-flash | 104.38 | 151.07 | -46.68s (-30.9%) | +| Gemini-2.5-flash-nothinking | 46.95 | 68.41 | -21.46s (-31.4%) | +| Qwen3-235b | 89.03 | 82.32 | +6.71s (+8.1%) | -# landing_page_generator: MCP vs A2A +# RecruitmentAssistant: MCP vs A2A -_Data not available for comparison_ +## Overall Summary (Averaged Across All Models) -# self_evaluation_loop_flow: MCP vs A2A +| MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | +| 444.91 | 501.47 | -56.56s (-11.3%) | -_Data not available for comparison_ +--- -# write_a_book_with_flows: MCP vs A2A +## Per-Model Comparison -_Data not available for comparison_ +| Model | MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 839.96 | 922.16 | -82.21s (-8.9%) | +| DeepSeek-V3-1 | 339.42 | 312.32 | +27.10s (+8.7%) | +| GPT-4o-mini | 129.31 | 178.62 | -49.31s (-27.6%) | +| GPT-5 | 59.50 | 67.32 | -7.81s (-11.6%) | +| Gemini-2.5-flash | 201.33 | 357.79 | -156.46s (-43.7%) | +| Gemini-2.5-flash-nothinking | 163.09 | 195.19 | -32.10s (-16.4%) | +| Qwen3-235b | 1381.73 | 1476.89 | -95.16s (-6.4%) | + +# LandingPageGenerator: MCP vs A2A + +## Overall Summary (Averaged Across All Models) + +| MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | +| 110.15 | 138.60 | -28.44s (-20.5%) | + +--- + +## Per-Model Comparison + +| Model | MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 274.53 | 302.33 | -27.80s (-9.2%) | +| DeepSeek-V3-1 | 135.84 | 215.52 | -79.68s (-37.0%) | +| GPT-4o-mini | 55.93 | 44.98 | +10.95s (+24.4%) | +| GPT-5 | 23.98 | 23.26 | +0.72s (+3.1%) | +| Gemini-2.5-flash | 91.42 | 76.95 | +14.46s (+18.8%) | +| Gemini-2.5-flash-nothinking | 97.29 | 98.14 | -0.85s (-0.9%) | +| Qwen3-235b | 92.07 | 208.99 | -116.92s (-55.9%) | + +# SocialMediaManager: MCP vs A2A + +## Overall Summary (Averaged Across All Models) + +| MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | +| 194.77 | 155.30 | +39.47s (+25.4%) | + +--- + +## Per-Model Comparison + +| Model | MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 735.24 | 557.94 | +177.30s (+31.8%) | +| DeepSeek-V3-1 | 219.42 | 158.60 | +60.81s (+38.3%) | +| GPT-4o-mini | 144.00 | 87.89 | +56.11s (+63.8%) | +| GPT-5 | 39.59 | 41.51 | -1.92s (-4.6%) | +| Gemini-2.5-flash | 59.94 | 65.45 | -5.51s (-8.4%) | +| Gemini-2.5-flash-nothinking | 38.63 | 40.70 | -2.07s (-5.1%) | +| Qwen3-235b | 126.60 | 135.01 | -8.42s (-6.2%) | + +# BookWriter: MCP vs A2A + +## Overall Summary (Averaged Across All Models) + +| MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | +| 335.67 | 361.87 | -26.20s (-7.2%) | + +--- + +## Per-Model Comparison + +| Model | MCP Mean (s) | A2A Mean (s) | Diff (MCP-A2A) | +| --- | --- | --- | --- | +| DeepSeek-R1 | 960.29 | 961.15 | -0.86s (-0.1%) | +| DeepSeek-V3-1 | 395.67 | 443.80 | -48.13s (-10.8%) | +| GPT-4o-mini | 112.23 | 101.15 | +11.08s (+11.0%) | +| GPT-5 | 140.17 | 193.58 | -53.41s (-27.6%) | +| Gemini-2.5-flash | 183.47 | 240.55 | -57.08s (-23.7%) | +| Gemini-2.5-flash-nothinking | 115.47 | 133.91 | -18.44s (-13.8%) | +| Qwen3-235b | 442.38 | 458.96 | -16.58s (-3.6%) | diff --git a/data/processed/RQ2/ECDFs/ecdf_BookWriter-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_BookWriter-A2A.pdf index 3fce6173341823974c00f1feaea660e263ce97e2..5fceb34cdee177f08fa11187a4a0a5ce2e41a257 100644 --- a/data/processed/RQ2/ECDFs/ecdf_BookWriter-A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_BookWriter-A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:9e599015731af91caa01cb06705f2f13a18e00e7cb525f96b8d1304125074130 +oid sha256:46c9971b6541ca196a33b2bc82d5faf18f83d8c8b59f43b96a0e9ffb3ff22839 size 15954 diff --git a/data/processed/RQ2/ECDFs/ecdf_BookWriter-H_A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_BookWriter-H_A2A.pdf index 373be678652a56bc361c2b98451bdb4eedcd1433..1adcab6c14a941ed92e1edfcc9a0946866290c1c 100644 --- a/data/processed/RQ2/ECDFs/ecdf_BookWriter-H_A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_BookWriter-H_A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:7dedabe80003ce75782b8c6b37459cc71c28f1e5f7beb13c0db26391babbb717 -size 15678 +oid sha256:483a08b0763a40ba68ff1bf88588d0e48e8a3c6e68b998486d2527f4c19ab411 +size 15894 diff --git a/data/processed/RQ2/ECDFs/ecdf_BookWriter-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_BookWriter-MCP.pdf index 43d5dc14dc3d637530e3b03ef871c18068568ba4..5c452ed26491c7235ad16a8cf6b0bedddb14a965 100644 --- a/data/processed/RQ2/ECDFs/ecdf_BookWriter-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_BookWriter-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:a55f0fd0c28bf9259d04637b8779bc362e8423308d93621fd49853e3dab24f2a +oid sha256:76a34e48d5e704371f9ab928a0479634722c03e2f5eee08e408a21df8464f2b5 size 15914 diff --git a/data/processed/RQ2/ECDFs/ecdf_EmailResponder-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_EmailResponder-MCP.pdf index a6c2bbdc87a116cb9484969aa47c77b8ffd78278..788c4d9e5bc3821eab08751fc3dcb470209181b0 100644 --- a/data/processed/RQ2/ECDFs/ecdf_EmailResponder-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_EmailResponder-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:13b4be4908f528b053596d93b6286a11bd45d08dcff2566e7929fa0c5e0cf320 +oid sha256:fb865f23aac0f997bd9ca900ef2cd72994062d4c7ae3fbd8edc117b1b33bc61c size 17728 diff --git a/data/processed/RQ2/ECDFs/ecdf_EmailResponder.pdf b/data/processed/RQ2/ECDFs/ecdf_EmailResponder.pdf index a67c7aa9388e92156e2a2fbb76add47c098ca7bd..8a6f39cad1772158b1c09e2e27910a5239b5c779 100644 --- a/data/processed/RQ2/ECDFs/ecdf_EmailResponder.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_EmailResponder.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:edd2b681e0acade6861421a91fbb43e3ec771c9e133c2fe357f849e3def13e6f +oid sha256:deae94f71028e064d08b21fb6b47f8a9f535b6eaf248a7e760dcb09e520befe6 size 18100 diff --git a/data/processed/RQ2/ECDFs/ecdf_GameBuilder-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_GameBuilder-MCP.pdf index 75f0ba18a6035f42e6f70ac11e334ce5f2d29d9f..92ffb2fea137d858fad28f68c232b22f3f0612e1 100644 --- a/data/processed/RQ2/ECDFs/ecdf_GameBuilder-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_GameBuilder-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:96737c854cb502b018c1c6e7ee3ae8554a7ebb2de60bcc988b46b6c0409ff179 +oid sha256:8b83f764e1be4cb9638ae1d2b7af251d4ac20babab51fba866c9d9b5ff831904 size 14307 diff --git a/data/processed/RQ2/ECDFs/ecdf_GameBuilder.pdf b/data/processed/RQ2/ECDFs/ecdf_GameBuilder.pdf index 2da21db45ccd1b4f0d55e06a2b961369079f3ddd..0c4e1758e24b27ebd8ec2fb1a2f47f0eeb94f921 100644 --- a/data/processed/RQ2/ECDFs/ecdf_GameBuilder.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_GameBuilder.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:bff1aae33966796ff95855f397daddec1f8271e1527a889afae46807f9364936 +oid sha256:4ef1b7ccdbcc12974093bf0166d4ae70546cfb905d381e01345ef038888a16ad size 14316 diff --git a/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-A2A.pdf index aa0233cb78aeedfceb2c5b0fa4abe2272b380a87..e5a526f8728adac89b4c0f920a5448772c42d27c 100644 --- a/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:002096ae37f1ab1e27e2ca32f2d240f5063e629e74f50531a9da85099423029a -size 15565 +oid sha256:a9d862ee93cfba0cff14b08994ead139d97398e3c7faa481d1507bf991016a57 +size 15534 diff --git a/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-H_A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-H_A2A.pdf index 8e45507effbed5ed5f14c70c9ea26f7e881df30d..eb977c06c771318e06af187ce2956ca1b13cb3be 100644 --- a/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-H_A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-H_A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:4497de82f43160357d4dc6b6481043d6e15b6146a9236e5764c0e0700d0dcd2a +oid sha256:6b358559934fd8a82575d492981c4c3ec38b4e408c9decfcd4411c2aa8a7a2b1 size 15558 diff --git a/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-MCP.pdf index 9fe514c281549732e1530d9ab20c54a6b9d023ee..3de86d0657c6ae53422efba5392a4de710519914 100644 --- a/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_LandingPageGenerator-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:6f673e6047f88a2ae74ad11af89f8ff461242c53c58db80943e7afe826013f5b -size 15545 +oid sha256:955c76c8d5a1948fd988216d808a980e3d0d8f039b2fe8dc872557dc10ed1cd5 +size 15530 diff --git a/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator-MCP.pdf index 55ec8b00a9046a930868020b7b416dbb48ab53e2..a9c87db697cbb7552acecf2b40df88feb97d716e 100644 --- a/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d2192a36927d1330de4935a761fbbed010826d9a79b8925fdc4b67a60b6c2d08 +oid sha256:2748c316d359675eb468faf24c0d7fea7188c6c9db4eb916007314b65157f9e5 size 15622 diff --git a/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator.pdf b/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator.pdf index 532ac0e1fc82519d7c7bf88d74e9a94e484ccd3d..7859337dce135da2bbfabec7e993cf15c8b8accb 100644 --- a/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_MarkdownValidator.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:261264c47eec6b6365277688ead535e582a6bb6cf45c6814505ce4d211342afe +oid sha256:050759c3b0eed5e4cf1ea33e4a40660e01b0ea18b3f3fd8f70a6ed1ecdb3b521 size 15600 diff --git a/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-A2A.pdf index f7dbdffeb074ae2dccc3f3fe48f2bca0b2cc92e9..7a5cf09b7b9b494ed9a273c6092b2b04c55b45ee 100644 --- a/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:7f914aa1bf3bde05a8a5ae0427358f435e28cbfd0f304ad4a06a2f51327b50df +oid sha256:cb5f61395cb8f56894e60f1b89341c9c4039a5e6b75d427fb376d24e40436144 size 15494 diff --git a/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-H_A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-H_A2A.pdf index f55de1386a199b8082b2151df26ebe27b28c0763..0f453c69cf0f1c2351cb893f01799af5b2e188a5 100644 --- a/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-H_A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-H_A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:0559368f6a2a72c4c68150cbda4bd4807fc107d4f6451045852a728d09e83410 -size 15968 +oid sha256:73671ece7dd2ef0845d6b7c3d3a8145ffea51fad28d83aaad3d0df169b7d438e +size 15428 diff --git a/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-MCP.pdf index 5d79ac3756515ce43b1132e7554350689ee23871..0d26dbe92a5875e9daabfb33c84cffd63122ed8f 100644 --- a/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_RecruitmentAssistant-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:3ebd55b54cf9dbb16d1d217fc4115e4d01f73bad5a3ddb0f61efa7f7720e3c58 +oid sha256:4dbf64fc5f86f3a9c99514499bd78b19a3ff7040e579dda50f6a2a4f49115d3a size 15437 diff --git a/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-A2A.pdf index 21eaf2e0638bec46ea70cf915456220e5a0c1124..21b24474b91f4f86c545c82813dc82ac778e68b1 100644 --- a/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:84c04b20af6af173a0d744418e57cb34bf39d31e697a695926a6bba7024a8577 +oid sha256:81193d359200c92961de77cb7f9d993114482b9ed3c334a780d0b869c7208cfd size 15860 diff --git a/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-H_A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-H_A2A.pdf index 460ef429f8b5c2e4d20ee2e90cca0241e1306306..05eb27c06448a60cd35d4934a40cdfa49062f8fd 100644 --- a/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-H_A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-H_A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:06aa657eef18bccf07ddf11d8c422efcae1d9a86b92c9a538dd5d6bb15cd7d47 -size 15804 +oid sha256:c8f15bd7743a84918d8358f330912bdc2fb5f030edd91af71dd7dded76894e90 +size 15787 diff --git a/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-MCP.pdf index 54f44a99fc8bf1895f357d3dc2fdf3f155c50e31..9a0e998f267641851a155ade632c8ffada28e5e5 100644 --- a/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_SQLAssistant-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fe2e3c28dbdff32f6c02d54d65fada8e2cd6585f7799173d484ee2c333202723 +oid sha256:0f9a870e9c85085ff3bf3edd1bfcb7f1c81dd8cc498c43a3dc0ca94fe4c734ae size 15864 diff --git a/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-A2A.pdf deleted file mode 100644 index 1539e44a3187782ea85a32d6930cc925d4acdea8..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-A2A.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:e48720e930dafd0d2d06dc1943760bc48671abd92218d8b88361a5e3155e952f -size 15860 diff --git a/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-A2A_mix.pdf b/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-A2A_mix.pdf deleted file mode 100644 index 58397be83d9a13768523ab672df249d6e0f4a2e3..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-A2A_mix.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:78d57688e23a05422efbb800600a70150e310baee4fd2872339763d8bc834fab -size 15787 diff --git a/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-MCP.pdf deleted file mode 100644 index 29b6b5f40050d45b97ec35bbb1b827ee7caa9d05..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_SQL_assistant-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:7ac8e8830a560f8efddb322fff4daab4065a9238a123310890f897a8eb9000b0 -size 15864 diff --git a/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-A2A.pdf index 7cb9da9706713084aa49363d6dbcdcea54dd0b33..89ae5d29e76e65f172a0cbdee6970cb402ec7442 100644 --- a/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ef6b27e08e2dfc6a9b46d024ebac91aa84e1f42cdc8f59fefe8f4345015c8b14 +oid sha256:6f74f9b9b5e69b82609909fc321d768abbd99811949225286e19ee8a56825c25 size 15841 diff --git a/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-H_A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-H_A2A.pdf index 3b36288f2d3f47e08903127fa480522694008c57..e0e44b498e26ba5bfcc95cb1d3e079c06aa3cdc2 100644 --- a/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-H_A2A.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-H_A2A.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:319681314f357a43c833a389728596c8f91a8ce89f2b08d7a5315a21b2a79d06 -size 15546 +oid sha256:9b509e7d21b81488694e0063f8d6be83836e022f387bb51bbe16dd85916066b2 +size 15805 diff --git a/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-MCP.pdf index 3ddc261168061874c7d85d275eba373d3a4ec189..efa5bdaaf5d5cede7990eda947d9e7b4d538e1c5 100644 --- a/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-MCP.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_SocialMediaManager-MCP.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:4660689cd443555771565b92b0bb925eeaf13790f8c86fc592e41b065731c7af +oid sha256:7e5ff0ee406bb8f7c667f4128f0eef53a5e1feedb5d7b1f12e714e0af88d698f size 15821 diff --git a/data/processed/RQ2/ECDFs/ecdf_email_auto_responder_flow-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_email_auto_responder_flow-MCP.pdf deleted file mode 100644 index bb874463b25f3fbc4c37024679b68d2ec862bec2..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_email_auto_responder_flow-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:4074778d32214d1484a94f027a7dac10f3c54a4f2b5f76e419248ee759887779 -size 17728 diff --git a/data/processed/RQ2/ECDFs/ecdf_email_auto_responder_flow.pdf b/data/processed/RQ2/ECDFs/ecdf_email_auto_responder_flow.pdf deleted file mode 100644 index a2a29d95151fb44463de94fccbdd244e0969228d..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_email_auto_responder_flow.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:f530cae645ea2c0f4164ca3d1aadc0129f7307f05492bcce61bc40f04d8034ef -size 18100 diff --git a/data/processed/RQ2/ECDFs/ecdf_game_builder-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_game_builder-MCP.pdf deleted file mode 100644 index ba4587f483a8a681a93b52de651ae2e0b867b04b..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_game_builder-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:023de9cbb80dbe9271c4a92c1ed5130c069ce420aa4bc9cd4839fe50f31e7ea0 -size 14307 diff --git a/data/processed/RQ2/ECDFs/ecdf_game_builder.pdf b/data/processed/RQ2/ECDFs/ecdf_game_builder.pdf deleted file mode 100644 index 076d70874b1a666df7f40b3572d448546dfb7f26..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_game_builder.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:924fe30dfa80134d8496f8c1ac1b4df91d90487bb363109f493edea722b874a9 -size 14316 diff --git a/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-A2A.pdf deleted file mode 100644 index 4bfb03926ed33cc76f45033d2ad5686c3382ee48..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-A2A.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:67155a12e49eca591b5a8af644ce37e342653148ceba2daacb48fc77117199fa -size 15494 diff --git a/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-A2A_mix.pdf b/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-A2A_mix.pdf deleted file mode 100644 index c0658010df0c6b31866ffeb4ec521163d2b8b4c3..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-A2A_mix.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:52898c0b13a50e90e2f3bc1092ba232ccc0c30ffe1d1bf31fb2d8bda9943297d -size 15428 diff --git a/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-MCP.pdf deleted file mode 100644 index feddd5fdb814c25dc4a3974f41c09c91aa2836dd..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_intelligent_recruitment_platform-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:24dc8a6bc14324e79a069c2a21f4f052d222ede91e88fc97114ab19c26bfcab4 -size 15437 diff --git a/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-A2A.pdf deleted file mode 100644 index 451d71fd4407ea7d35df833a19cf0f4f8ccb4423..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-A2A.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:05f95d74c35d558643d0caa1542c2cc2a19d11488396dd2a61f2df79ad5339f7 -size 15534 diff --git a/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-A2A_mix.pdf b/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-A2A_mix.pdf deleted file mode 100644 index 0426d9538b66a4ea7c1700988edee06332241cdf..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-A2A_mix.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:458c2865084a4c628515f2d6f0d462cd5ccaa65b1999f2d8134dfea47ef7d957 -size 15558 diff --git a/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-MCP.pdf deleted file mode 100644 index 08d7c8287e57d3f1d4a31b730eefbc19e6949a0d..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_landing_page_generator-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:25a40dd77393830eb5cfe5e592c667dc631ee01f6e32aa3914317a867d8eb5e7 -size 15530 diff --git a/data/processed/RQ2/ECDFs/ecdf_markdown_validator-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_markdown_validator-MCP.pdf deleted file mode 100644 index 06484143bc716a973d987a0d7345de55c8698549..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_markdown_validator-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:94113d90a38c154f1225631c7e7be857bc7a4f7add443f1de64209a27a7f77f6 -size 15622 diff --git a/data/processed/RQ2/ECDFs/ecdf_markdown_validator.pdf b/data/processed/RQ2/ECDFs/ecdf_markdown_validator.pdf deleted file mode 100644 index 40c626cb2aa76522033d5ee910f5d0e4d7ea792d..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_markdown_validator.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:061001684d9138211106bb8a00d82e05ca2e86e13dba2815c5abcb07cf6f7d8e -size 15600 diff --git a/data/processed/RQ2/ECDFs/ecdf_overall_time_weighted.pdf b/data/processed/RQ2/ECDFs/ecdf_overall_time_weighted.pdf index 0a42433cc4e79ebbfcd4f4898e082455bc328a3e..3dbe2a69312f5b899a24f38f9a506ae4ab5dcf1d 100644 --- a/data/processed/RQ2/ECDFs/ecdf_overall_time_weighted.pdf +++ b/data/processed/RQ2/ECDFs/ecdf_overall_time_weighted.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:9ac041779d047f9d0754e8aab9101e259a83571456d0df50adf34942c96dc759 +oid sha256:84a41197a2a3f91cf850fc1dcc08edb4dacf2e00d4247ba98db6d66337c2a423 size 28806 diff --git a/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-A2A.pdf deleted file mode 100644 index c511e915102cd4a730d4f9d6de154f4cc9f6ef39..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-A2A.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:b040db491f7c3028160ea19da4d6c3e9ed381837137c6042d3d83c0fc176eea4 -size 15841 diff --git a/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-A2A_mix.pdf b/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-A2A_mix.pdf deleted file mode 100644 index 16baff0c5bb2ce41e25b1d70681148c74b16822b..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-A2A_mix.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:4b1921a7d5421b254549c6637ab9db083ddf551fbd24b482e3b782403598233e -size 15805 diff --git a/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-MCP.pdf deleted file mode 100644 index e94031ed16538fdadbe47210110e0c535778dba5..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_self_evaluation_loop_flow-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:eac36351e32631cec9a2e32778248a22ffc8526b507905a4dbce42eff64fb1f5 -size 15821 diff --git a/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-A2A.pdf b/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-A2A.pdf deleted file mode 100644 index cbad3397d1f0e3a1a095bcbfb264722d08972e67..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-A2A.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:564943ff1a48b56edeeed0852e56bdc6fdfa2e7dc42e093231405f51db96aeca -size 15954 diff --git a/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-A2A_mix.pdf b/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-A2A_mix.pdf deleted file mode 100644 index 577d9453794740bb7b3f0f74820f1a8a8c874252..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-A2A_mix.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:4107343b9c66ec22ff521ea812aecf4fefea101485aa6ddb72801d3e0797ff61 -size 15894 diff --git a/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-MCP.pdf b/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-MCP.pdf deleted file mode 100644 index 9ed34cb4a581d12188269854f73305ca3822379e..0000000000000000000000000000000000000000 --- a/data/processed/RQ2/ECDFs/ecdf_write_a_book_with_flows-MCP.pdf +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:8fc264a8617b567629155d82f3ccea1f2019e8d4942473b051744f470d41dec4 -size 15914 diff --git a/data/processed/RQ2/plot_total_classified_ecdf.py b/data/processed/RQ2/plot_total_classified_ecdf.py index b3ed93c5bbe63183288473d2aa62dd817f955506..04aa41a00b2035d7c46f111e208f8087fcf9872f 100644 --- a/data/processed/RQ2/plot_total_classified_ecdf.py +++ b/data/processed/RQ2/plot_total_classified_ecdf.py @@ -546,8 +546,8 @@ def generate_version_overall_comparison( def get_base_project_name(scenario: str) -> str: - if scenario.endswith("-A2A_mix"): - return scenario[: -len("-A2A_mix")] + if scenario.endswith("-H_A2A"): + return scenario[: -len("-H_A2A")] if scenario.endswith("-A2A"): return scenario[: -len("-A2A")] if scenario.endswith("-MCP"): @@ -589,15 +589,12 @@ def generate_overall_model_comparison( # Summary table lines.append("### Model Statistics Summary\n\n") - lines.append("| Model | Mean (s) | Median (s) | Count | Total (s) |\n") - lines.append("| --- | --- | --- | --- | --- |\n") + lines.append("| Model | Mean (s) |\n") + lines.append("| --- | --- |\n") for model in all_models: stats = model_stats[model] - lines.append( - f"| {model} | {stats['mean']:.2f} | {stats['median']:.2f} | " - f"{stats['count']} | {stats['total']:.2f} |\n" - ) + lines.append(f"| {model} | {stats['mean']:.2f} |\n") # Relative performance vs fastest (only if at least two models with data) positive_models = [m for m in all_models if model_stats[m]["mean"] > 0] @@ -630,7 +627,7 @@ def generate_overall_model_comparison( slowest_slowdown = (slowest_mean - fastest_mean) / fastest_mean * 100 lines.append( f"\n**Slowest model:** {slowest_model} ({slowest_mean:.2f}s mean, " - f"{slowest_slowdown:+.1f}% slower than baseline)\n" + f"{slowest_slowdown:+.1f}% slower than baseline {fastest_model})\n" ) lines.append("\n") @@ -666,15 +663,12 @@ def generate_project_model_comparison( # Summary table lines.append("## Model Statistics Summary\n\n") - lines.append("| Model | Mean (s) | Median (s) | Count | Total (s) |\n") - lines.append("| --- | --- | --- | --- | --- |\n") + lines.append("| Model | Mean (s) |\n") + lines.append("| --- | --- |\n") for model in all_models: stats = model_stats[model] - lines.append( - f"| {model} | {stats['mean']:.2f} | {stats['median']:.2f} | " - f"{stats['count']} | {stats['total']:.2f} |\n" - ) + lines.append(f"| {model} | {stats['mean']:.2f} |\n") # Pairwise comparison: compare each model against the fastest one if len(all_models) > 1: @@ -715,7 +709,7 @@ def generate_project_model_comparison( slowest_slowdown = (slowest_mean - fastest_mean) / fastest_mean * 100 lines.append( f"\n**Slowest model:** {slowest_model} ({slowest_mean:.2f}s mean, " - f"{slowest_slowdown:+.1f}% slower than baseline)\n" + f"{slowest_slowdown:+.1f}% slower than baseline {fastest_model})\n" ) lines.append("\n") @@ -887,42 +881,37 @@ def main() -> None: print("Generating time comparison summaries...") print("=" * 60) - # 1. MCP vs Hardcoded comparisons mcp_hardcoded_projects = [ "MarkdownValidator", "GameBuilder", "EmailResponder", ] + version_projects = [ + "SQLAssistant", + "RecruitmentAssistant", + "LandingPageGenerator", + "SocialMediaManager", + "BookWriter", + ] + + # 1. MCP vs Hardcoded comparisons comparison_lines = [] - # Add overall summary first comparison_lines.append( generate_mcp_vs_hardcoded_overall_comparison( mcp_hardcoded_projects, scenario_time_data ) ) - - # Then add per-project comparisons for project in mcp_hardcoded_projects: comparison_lines.append( generate_mcp_vs_hardcoded_comparison(project, scenario_time_data) ) - comparison_md_path = out_dir / "Time_Comparison_MCP_vs_Hardcoded.md" comparison_md_path.write_text("".join(comparison_lines), encoding="utf-8") print(f"Created: {comparison_md_path}") # 2. MCP vs A2A comparisons - version_projects = [ - "SQL_assistant", - "intelligent_recruitment_platform", - "landing_page_generator", - "self_evaluation_loop_flow", - "write_a_book_with_flows", - ] - comparison_lines = [] - # Add overall summary first comparison_lines.append( generate_version_overall_comparison( version_projects, @@ -934,54 +923,47 @@ def main() -> None: "MCP vs A2A Time Comparison", ) ) - - # Then add per-project comparisons for project in version_projects: comparison_lines.append( generate_version_comparison( project, "-MCP", "-A2A", scenario_time_data, "MCP", "A2A" ) ) - comparison_md_path = out_dir / "Time_Comparison_MCP_vs_A2A.md" comparison_md_path.write_text("".join(comparison_lines), encoding="utf-8") print(f"Created: {comparison_md_path}") - # 3. A2A vs A2A_mix comparisons + # 3. A2A vs H_A2A comparisons comparison_lines = [] - # Add overall summary first comparison_lines.append( generate_version_overall_comparison( version_projects, "-A2A", - "-A2A_mix", + "-H_A2A", scenario_time_data, "A2A", - "A2A_mix", - "A2A vs A2A_mix Time Comparison", + "H_A2A", + "A2A vs H_A2A Time Comparison", ) ) - - # Then add per-project comparisons for project in version_projects: comparison_lines.append( generate_version_comparison( - project, "-A2A", "-A2A_mix", scenario_time_data, "A2A", "A2A_mix" + project, "-A2A", "-H_A2A", scenario_time_data, "A2A", "H_A2A" ) ) - - comparison_md_path = out_dir / "Time_Comparison_A2A_vs_A2A_mix.md" + comparison_md_path = out_dir / "Time_Comparison_A2A_vs_H_A2A.md" comparison_md_path.write_text("".join(comparison_lines), encoding="utf-8") print(f"Created: {comparison_md_path}") - # 4. Generate detailed A2A vs A2A_mix comparison for each project - print("\nGenerating detailed A2A vs A2A_mix per-project comparisons...") + # 4. Generate detailed A2A vs H_A2A comparison for each project + print("\nGenerating detailed A2A vs H_A2A per-project comparisons...") a2a_mix_comparison_lines = [] a2a_mix_comparison_lines.append( - "# A2A vs A2A_mix: Detailed Per-Project Comparison\n\n" + "# A2A vs H_A2A: Detailed Per-Project Comparison\n\n" ) a2a_mix_comparison_lines.append( - "This document compares A2A and A2A_mix architectures for each project, " + "This document compares A2A and H_A2A architectures for each project, " ) a2a_mix_comparison_lines.append( "showing both per-model and overall statistics.\n\n" @@ -999,7 +981,7 @@ def main() -> None: for project in version_projects: scenario_a2a = f"{project}-A2A" - scenario_a2a_mix = f"{project}-A2A_mix" + scenario_a2a_mix = f"{project}-H_A2A" if ( scenario_a2a not in scenario_time_data @@ -1040,7 +1022,7 @@ def main() -> None: pct = (diff / avg_a2a * 100) if avg_a2a > 0 else 0 a2a_mix_comparison_lines.append("### Overall Summary\n\n") a2a_mix_comparison_lines.append( - "| A2A Mean (s) | A2A_mix Mean (s) | Diff (A2A_mix - A2A) |\n" + "| A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) |\n" ) a2a_mix_comparison_lines.append("| --- | --- | --- |\n") a2a_mix_comparison_lines.append( @@ -1059,7 +1041,7 @@ def main() -> None: # Per-model comparison a2a_mix_comparison_lines.append("### Per-Model Comparison\n\n") a2a_mix_comparison_lines.append( - "| Model | A2A Mean (s) | A2A_mix Mean (s) | Diff (A2A_mix - A2A) |\n" + "| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) |\n" ) a2a_mix_comparison_lines.append("| --- | --- | --- | --- |\n") @@ -1082,7 +1064,7 @@ def main() -> None: a2a_mix_comparison_lines.insert( 4, "## Overall (All Projects)\n\n" - "| Project | A2A Mean (s) | A2A_mix Mean (s) | Diff (A2A_mix - A2A) |\n" + "| Project | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) |\n" "| --- | --- | --- | --- |\n" + "".join( f"| {d['project']} | {d['a2a']:.2f} | {d['mix']:.2f} | {d['diff']:+.2f}s ({d['pct']:+.1f}%) |\n" @@ -1095,7 +1077,7 @@ def main() -> None: per_model_lines = [] per_model_lines.append("## All Projects Combined (Per-Model)\n\n") per_model_lines.append( - "| Model | A2A Mean (s) | A2A_mix Mean (s) | Diff (A2A_mix - A2A) |\n" + "| Model | A2A Mean (s) | H_A2A Mean (s) | Diff (H_A2A - A2A) |\n" ) per_model_lines.append("| --- | --- | --- | --- |\n") for model in MODEL_ORDER: @@ -1117,7 +1099,7 @@ def main() -> None: # insert after top intro (after first 4 elements added earlier) a2a_mix_comparison_lines[5:5] = per_model_lines - a2a_mix_md_path = out_dir / "A2A_vs_A2A_mix_Detailed_Comparison.md" + a2a_mix_md_path = out_dir / "A2A_vs_H_A2A_Detailed_Comparison.md" a2a_mix_md_path.write_text("".join(a2a_mix_comparison_lines), encoding="utf-8") print(f"Created: {a2a_mix_md_path}") @@ -1131,9 +1113,7 @@ def main() -> None: f"This document compares model performance within each of the {len(all_project_names)} projects.\n\n" ) model_comparison_lines.append("Each project shows:\n") - model_comparison_lines.append( - "- Model statistics (mean, median, count, total time)\n" - ) + model_comparison_lines.append("- Model statistics (mean)\n") model_comparison_lines.append( "- Relative performance compared to the fastest model\n\n" ) diff --git a/data/processed/RQ3/Violin/SQL_assistant-A2A_mix_total_tokens_violin.pdf b/data/processed/RQ3/Violin/SQL_assistant-A2A_mix_total_tokens_violin.pdf index a9a1be681bf900b9200da68f000f2dc708f158ec..c17a513075ab9c9ed3ce363d04232749c4b70ce5 100644 --- a/data/processed/RQ3/Violin/SQL_assistant-A2A_mix_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/SQL_assistant-A2A_mix_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:1775639ea13c20095110b5cb0bb86a010eca6b21ef283dd3734ea9d0b5e59ff6 -size 42739 +oid sha256:e2ef4ee6fd871b76950bf50acbdb1d39ee8bce23486e4b313d5eb2dcc7f0d4b9 +size 42992 diff --git a/data/processed/RQ3/Violin/SQL_assistant-A2A_total_tokens_violin.pdf b/data/processed/RQ3/Violin/SQL_assistant-A2A_total_tokens_violin.pdf index 485a6ae1981060b025715c7ddf85ce58f1e23ee9..05a008d0e40d13fb294fc27ab38f4eb342564a81 100644 --- a/data/processed/RQ3/Violin/SQL_assistant-A2A_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/SQL_assistant-A2A_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fde2aff67729d7fd440ff721219f6278eea7dc76a079125819f8f143cdc0cfce -size 41937 +oid sha256:7368dfb27c83a38dd0f1e18178c5b0c397d0a37c1b0985173cc86f488b280afb +size 42192 diff --git a/data/processed/RQ3/Violin/SQL_assistant-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/SQL_assistant-MCP_total_tokens_violin.pdf index fbfe243626aed47a1ccf5c3d20cbd1f10ab4bb89..654675e42d0d2f4fb90715ac72978aca2c447833 100644 --- a/data/processed/RQ3/Violin/SQL_assistant-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/SQL_assistant-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fc73433a6cea10136db92cb8ca8268a1fc0379947ca693459dac6e8659158e4f -size 42858 +oid sha256:10c0be8f3528dfbfabf238b58c73bca8670dcc43255909a806b2a8d131ff51e9 +size 43109 diff --git a/data/processed/RQ3/Violin/architecture_token_deltas_a2a_to_h-a2a.md b/data/processed/RQ3/Violin/architecture_token_deltas_a2a_to_h-a2a.md new file mode 100644 index 0000000000000000000000000000000000000000..e8df2804582881f85b568570237c8800e7589693 --- /dev/null +++ b/data/processed/RQ3/Violin/architecture_token_deltas_a2a_to_h-a2a.md @@ -0,0 +1,95 @@ +# Token shifts across architectures + +Series: **A2A → H-A2A** + +Each table shows the absolute change (Δ) and the relative percentage change of the mean total tokens. + +## BookWriter + +| Model | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| GPT-5 | 181,022 | 117,358 | -63,664 (-35.2%) | +| GPT-4o-mini | 139,665 | 102,564 | -37,101 (-26.6%) | +| DeepSeek-V3.1 | 272,275 | 141,204 | -131,071 (-48.1%) | +| DeepSeek-R1 | 186,678 | 129,883 | -56,795 (-30.4%) | +| Gemini-2.5 | 110,126 | 91,585 | -18,541 (-16.8%) | +| Gemini-2.5-NT | 86,249 | 65,826 | -20,422 (-23.7%) | +| Qwen3-235b | 118,406 | 76,512 | -41,893 (-35.4%) | + +Project-level average (all models combined) + +| Metric | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| Avg tokens (all models) | 156,346 | 103,562 | -52,784 (-33.8%) | + +## LandingPageGenerator + +| Model | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| GPT-5 | 14,965 | 25,296 | +10,331 (69.0%) | +| GPT-4o-mini | 20,340 | 43,589 | +23,249 (114.3%) | +| DeepSeek-V3.1 | 28,814 | 34,496 | +5,682 (19.7%) | +| DeepSeek-R1 | 28,657 | 57,203 | +28,546 (99.6%) | +| Gemini-2.5 | 21,256 | 45,250 | +23,995 (112.9%) | +| Gemini-2.5-NT | 22,826 | 14,949 | -7,876 (-34.5%) | +| Qwen3-235b | 18,142 | 17,628 | -514 (-2.8%) | + +Project-level average (all models combined) + +| Metric | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| Avg tokens (all models) | 22,143 | 34,059 | +11,916 (53.8%) | + +## RecruitmentAssistant + +| Model | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| GPT-5 | 31,991 | 32,672 | +681 (2.1%) | +| GPT-4o-mini | 74,951 | 101,936 | +26,985 (36.0%) | +| DeepSeek-V3.1 | 65,662 | 65,048 | -614 (-0.9%) | +| DeepSeek-R1 | 72,968 | 79,028 | +6,060 (8.3%) | +| Gemini-2.5 | 68,719 | 57,128 | -11,591 (-16.9%) | +| Gemini-2.5-NT | 55,660 | 36,887 | -18,772 (-33.7%) | +| Qwen3-235b | 97,717 | 73,259 | -24,458 (-25.0%) | + +Project-level average (all models combined) + +| Metric | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| Avg tokens (all models) | 66,810 | 63,708 | -3,101 (-4.6%) | + +## SQLAssistant + +| Model | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| GPT-5 | 11,189 | 8,956 | -2,233 (-20.0%) | +| GPT-4o-mini | 18,635 | 13,734 | -4,901 (-26.3%) | +| DeepSeek-V3.1 | 36,805 | 16,479 | -20,326 (-55.2%) | +| DeepSeek-R1 | 96,469 | 69,364 | -27,105 (-28.1%) | +| Gemini-2.5 | 32,531 | 24,448 | -8,083 (-24.8%) | +| Gemini-2.5-NT | 25,782 | 18,476 | -7,306 (-28.3%) | +| Qwen3-235b | 17,381 | 14,779 | -2,602 (-15.0%) | + +Project-level average (all models combined) + +| Metric | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| Avg tokens (all models) | 34,113 | 23,748 | -10,365 (-30.4%) | + +## SocialMediaManager + +| Model | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| GPT-5 | 40,570 | 16,796 | -23,773 (-58.6%) | +| GPT-4o-mini | 97,814 | 43,322 | -54,492 (-55.7%) | +| DeepSeek-V3.1 | 87,367 | 57,449 | -29,918 (-34.2%) | +| DeepSeek-R1 | 124,960 | 98,225 | -26,735 (-21.4%) | +| Gemini-2.5 | 12,955 | 20,423 | +7,468 (57.6%) | +| Gemini-2.5-NT | 11,429 | 9,642 | -1,787 (-15.6%) | +| Qwen3-235b | 55,866 | 29,440 | -26,426 (-47.3%) | + +Project-level average (all models combined) + +| Metric | A2A | H-A2A | Δ H-A2A-A2A | +| --- | --- | --- | --- | +| Avg tokens (all models) | 61,566 | 39,328 | -22,237 (-36.1%) | \ No newline at end of file diff --git a/data/processed/RQ3/Violin/architecture_token_deltas_crewai_to_mcp.md b/data/processed/RQ3/Violin/architecture_token_deltas_crewai_to_mcp.md new file mode 100644 index 0000000000000000000000000000000000000000..0b59bbd3c80a8d2bf93d56dbc60c6205bcc290a7 --- /dev/null +++ b/data/processed/RQ3/Violin/architecture_token_deltas_crewai_to_mcp.md @@ -0,0 +1,59 @@ +# Token shifts across architectures + +Series: **Pure CrewAI → MCP** + +Each table shows the absolute change (Δ) and the relative percentage change of the mean total tokens. + +## EmailResponder + +| Model | Pure CrewAI | MCP | Δ MCP-Pure CrewAI | +| --- | --- | --- | --- | +| GPT-5 | 5,562 | 6,903 | +1,342 (24.1%) | +| GPT-4o-mini | 8,044 | 9,437 | +1,393 (17.3%) | +| DeepSeek-V3.1 | 7,431 | 9,580 | +2,149 (28.9%) | +| DeepSeek-R1 | 16,982 | 15,403 | -1,579 (-9.3%) | +| Gemini-2.5 | 35,189 | 38,066 | +2,877 (8.2%) | +| Gemini-2.5-NT | 42,421 | 35,133 | -7,287 (-17.2%) | +| Qwen3-235b | 5,900 | 7,434 | +1,533 (26.0%) | + +Project-level average (all models combined) + +| Metric | Pure CrewAI | MCP | Δ MCP-Pure CrewAI | +| --- | --- | --- | --- | +| Avg tokens (all models) | 17,361 | 17,422 | +61 (0.4%) | + +## GameBuilder + +| Model | Pure CrewAI | MCP | Δ MCP-Pure CrewAI | +| --- | --- | --- | --- | +| GPT-5 | 22,904 | 22,425 | -479 (-2.1%) | +| GPT-4o-mini | 17,823 | 47,996 | +30,173 (169.3%) | +| DeepSeek-V3.1 | 41,341 | 55,289 | +13,947 (33.7%) | +| DeepSeek-R1 | 44,536 | 44,851 | +316 (0.7%) | +| Gemini-2.5 | 72,772 | 75,519 | +2,747 (3.8%) | +| Gemini-2.5-NT | 67,964 | 50,782 | -17,181 (-25.3%) | +| Qwen3-235b | 24,212 | 24,039 | -173 (-0.7%) | + +Project-level average (all models combined) + +| Metric | Pure CrewAI | MCP | Δ MCP-Pure CrewAI | +| --- | --- | --- | --- | +| Avg tokens (all models) | 41,650 | 45,843 | +4,193 (10.1%) | + +## MarkdownValidator + +| Model | Pure CrewAI | MCP | Δ MCP-Pure CrewAI | +| --- | --- | --- | --- | +| GPT-5 | 1,853 | 1,967 | +115 (6.2%) | +| GPT-4o-mini | 1,853 | 1,942 | +88 (4.8%) | +| DeepSeek-V3.1 | 1,948 | 2,063 | +115 (5.9%) | +| DeepSeek-R1 | 5,912 | 5,885 | -26 (-0.4%) | +| Gemini-2.5 | 135,224 | 95,503 | -39,722 (-29.4%) | +| Gemini-2.5-NT | 31,330 | 61,899 | +30,569 (97.6%) | +| Qwen3-235b | 1,854 | 1,994 | +140 (7.6%) | + +Project-level average (all models combined) + +| Metric | Pure CrewAI | MCP | Δ MCP-Pure CrewAI | +| --- | --- | --- | --- | +| Avg tokens (all models) | 25,711 | 24,465 | -1,246 (-4.8%) | \ No newline at end of file diff --git a/data/processed/RQ3/Violin/architecture_token_deltas_mcp_to_a2a.md b/data/processed/RQ3/Violin/architecture_token_deltas_mcp_to_a2a.md new file mode 100644 index 0000000000000000000000000000000000000000..3098afac21c24c2f42dea95dd6e8862c4badea55 --- /dev/null +++ b/data/processed/RQ3/Violin/architecture_token_deltas_mcp_to_a2a.md @@ -0,0 +1,95 @@ +# Token shifts across architectures + +Series: **MCP → A2A** + +Each table shows the absolute change (Δ) and the relative percentage change of the mean total tokens. + +## BookWriter + +| Model | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| GPT-5 | 190,842 | 181,022 | -9,820 (-5.1%) | +| GPT-4o-mini | 133,339 | 139,665 | +6,326 (4.7%) | +| DeepSeek-V3.1 | 184,315 | 272,275 | +87,960 (47.7%) | +| DeepSeek-R1 | 190,705 | 186,678 | -4,027 (-2.1%) | +| Gemini-2.5 | 107,154 | 110,126 | +2,973 (2.8%) | +| Gemini-2.5-NT | 82,546 | 86,249 | +3,703 (4.5%) | +| Qwen3-235b | 110,716 | 118,406 | +7,690 (6.9%) | + +Project-level average (all models combined) + +| Metric | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| Avg tokens (all models) | 142,802 | 156,346 | +13,543 (9.5%) | + +## LandingPageGenerator + +| Model | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| GPT-5 | 12,653 | 14,965 | +2,312 (18.3%) | +| GPT-4o-mini | 19,250 | 20,340 | +1,090 (5.7%) | +| DeepSeek-V3.1 | 15,866 | 28,814 | +12,948 (81.6%) | +| DeepSeek-R1 | 28,371 | 28,657 | +286 (1.0%) | +| Gemini-2.5 | 26,527 | 21,256 | -5,271 (-19.9%) | +| Gemini-2.5-NT | 33,428 | 22,826 | -10,602 (-31.7%) | +| Qwen3-235b | 14,478 | 18,142 | +3,664 (25.3%) | + +Project-level average (all models combined) + +| Metric | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| Avg tokens (all models) | 21,510 | 22,143 | +632 (2.9%) | + +## RecruitmentAssistant + +| Model | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| GPT-5 | 32,411 | 31,991 | -419 (-1.3%) | +| GPT-4o-mini | 99,933 | 74,951 | -24,982 (-25.0%) | +| DeepSeek-V3.1 | 65,748 | 65,662 | -85 (-0.1%) | +| DeepSeek-R1 | 84,686 | 72,968 | -11,718 (-13.8%) | +| Gemini-2.5 | 64,864 | 68,719 | +3,855 (5.9%) | +| Gemini-2.5-NT | 54,402 | 55,660 | +1,257 (2.3%) | +| Qwen3-235b | 87,098 | 97,717 | +10,620 (12.2%) | + +Project-level average (all models combined) + +| Metric | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| Avg tokens (all models) | 69,877 | 66,810 | -3,068 (-4.4%) | + +## SQLAssistant + +| Model | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| GPT-5 | 13,881 | 11,189 | -2,692 (-19.4%) | +| GPT-4o-mini | 24,241 | 18,635 | -5,606 (-23.1%) | +| DeepSeek-V3.1 | 25,637 | 36,805 | +11,169 (43.6%) | +| DeepSeek-R1 | 80,522 | 96,469 | +15,947 (19.8%) | +| Gemini-2.5 | 19,110 | 32,531 | +13,422 (70.2%) | +| Gemini-2.5-NT | 15,014 | 25,782 | +10,769 (71.7%) | +| Qwen3-235b | 17,232 | 17,381 | +149 (0.9%) | + +Project-level average (all models combined) + +| Metric | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| Avg tokens (all models) | 27,948 | 34,113 | +6,165 (22.1%) | + +## SocialMediaManager + +| Model | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| GPT-5 | 38,940 | 40,570 | +1,630 (4.2%) | +| GPT-4o-mini | 108,089 | 97,814 | -10,275 (-9.5%) | +| DeepSeek-V3.1 | 95,721 | 87,367 | -8,355 (-8.7%) | +| DeepSeek-R1 | 109,146 | 124,960 | +15,814 (14.5%) | +| Gemini-2.5 | 12,749 | 12,955 | +207 (1.6%) | +| Gemini-2.5-NT | 11,480 | 11,429 | -51 (-0.4%) | +| Qwen3-235b | 46,557 | 55,866 | +9,308 (20.0%) | + +Project-level average (all models combined) + +| Metric | MCP | A2A | Δ A2A-MCP | +| --- | --- | --- | --- | +| Avg tokens (all models) | 60,383 | 61,566 | +1,183 (2.0%) | \ No newline at end of file diff --git a/data/processed/RQ3/Violin/email_auto_responder_flow-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/email_auto_responder_flow-MCP_total_tokens_violin.pdf index b2d63970274f164e6ad63a695f339536d9ea8f8e..21b23a5036bfea11eda43024c6575ef5bc048652 100644 --- a/data/processed/RQ3/Violin/email_auto_responder_flow-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/email_auto_responder_flow-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c52ff58ee4edfc6db866a6c8a185aa7cbb6ba12603399c9d911ae3c1410b186b -size 33564 +oid sha256:e4040c3c585d42d668220510ed4ace950bcfb2f2028519d65c63a2a21076f35b +size 33833 diff --git a/data/processed/RQ3/Violin/email_auto_responder_flow_total_tokens_violin.pdf b/data/processed/RQ3/Violin/email_auto_responder_flow_total_tokens_violin.pdf index 01ec6b7aa2716bf3e686bcf69f90e77e5ce057f2..9827d0a93294b20021fa6d8bc564ee87a8edf3f6 100644 --- a/data/processed/RQ3/Violin/email_auto_responder_flow_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/email_auto_responder_flow_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:5b8150e4364ba3e40b7d2c262fa8cd1b42268367e1b899015480a93a7191ae0a -size 34323 +oid sha256:b0b3d95d2584aaabe75ce28a34f50e3937aaeb6ad6267aff6203c70d63d227be +size 34591 diff --git a/data/processed/RQ3/Violin/game_builder-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/game_builder-MCP_total_tokens_violin.pdf index d6cd5c5ff6af697f4d0632830fcf91649556317d..846f5a4d2c6aefdc9902888c77a690193a75290e 100644 --- a/data/processed/RQ3/Violin/game_builder-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/game_builder-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:41515fccdd71f6ffee82eacd430c4b2598e698250887ebb5f750114268ae2f1d -size 27249 +oid sha256:c10c5d39bda006b2a0dd747969e2dd25a3038ef6687b993b2d52fc26bffc302f +size 27515 diff --git a/data/processed/RQ3/Violin/game_builder_total_tokens_violin.pdf b/data/processed/RQ3/Violin/game_builder_total_tokens_violin.pdf index 864dc9058e82f1289279c04dcf5fbee467939b71..9ee6dae45c52aa0605d4a448e3d705a7bdbe3474 100644 --- a/data/processed/RQ3/Violin/game_builder_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/game_builder_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ea28f24a09d8d8edd879ad289ae754e3edb838af1d252531e2185ca2f0fe0db9 -size 28886 +oid sha256:a39458209d12f197f47d327b6715982900ed9a8d135f71e1628636ba14240268 +size 29150 diff --git a/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_mix_total_tokens_violin.pdf b/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_mix_total_tokens_violin.pdf index f567a2687269cc30302ea4dbf23485aa119722c6..a0b759e39d9603cd17420f23e18a7ab7b4171225 100644 --- a/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_mix_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_mix_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d8aea7ce5422d8590e866562d3d3c26ea5b9163295aeb73b201c504315ddcb72 -size 31932 +oid sha256:8ae22bfa81a2d30992a028abc9ebe5923c025e80282469e9af5ffeaa0d6997de +size 32196 diff --git a/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_total_tokens_violin.pdf b/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_total_tokens_violin.pdf index 0011c546764055e9c8b40a41d76abf2fe1826017..78c477678e72253942935834f249d0cab2358263 100644 --- a/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/intelligent_recruitment_platform-A2A_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:345afb926ca6580d78a63927ea457b6383625e34854b6a074a22ad5c0ea80c1a -size 32136 +oid sha256:8eb80d9de4acbd6409a20f207ce5ad2f082bd54b3c61cadc5faf7132b5f5f8e2 +size 32400 diff --git a/data/processed/RQ3/Violin/intelligent_recruitment_platform-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/intelligent_recruitment_platform-MCP_total_tokens_violin.pdf index 1085828d4b7fd61375f99dea0e8852f08fe62a8c..380d536aa31d3242a1d376025ff63737a03966e6 100644 --- a/data/processed/RQ3/Violin/intelligent_recruitment_platform-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/intelligent_recruitment_platform-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:2e02ab5493c362d9f796aeabcc84fc0096ee2e49e0eab6f0eefd2eabba12239b -size 33065 +oid sha256:6b5489bae6fb5dfc4a42087586b95deae8f7ba693d1dbb558c8ef06f1b619ff5 +size 33330 diff --git a/data/processed/RQ3/Violin/landing_page_generator-A2A_mix_total_tokens_violin.pdf b/data/processed/RQ3/Violin/landing_page_generator-A2A_mix_total_tokens_violin.pdf index b2eae47c2115166a1fc10a408311b931fae6dceb..0b44a44da9953b846dde7fcd6769c5456a53e248 100644 --- a/data/processed/RQ3/Violin/landing_page_generator-A2A_mix_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/landing_page_generator-A2A_mix_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ef0e2a63879d2799c840b6d7fa65b7bdb5fc4e98bef19cf76e7f2fde392ce449 -size 35537 +oid sha256:1bdac78f49c2e294084b8546746958989235bdc4a06c105654a618388cb2b6fa +size 35800 diff --git a/data/processed/RQ3/Violin/landing_page_generator-A2A_total_tokens_violin.pdf b/data/processed/RQ3/Violin/landing_page_generator-A2A_total_tokens_violin.pdf index cb809e51ed472e93a3825ba45e9b21794cf1ef35..db0ee083d2db5f0547f51c13d65e589569380e58 100644 --- a/data/processed/RQ3/Violin/landing_page_generator-A2A_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/landing_page_generator-A2A_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:6888cd9b592c5dec801d3298f07d20d4108e8302e672a8ca9e872627b82a4e62 -size 35353 +oid sha256:39e9bbd84eb049866e9fa48d940f20f65dd7a5f8217e87e313fc6a5bbacf7067 +size 35619 diff --git a/data/processed/RQ3/Violin/landing_page_generator-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/landing_page_generator-MCP_total_tokens_violin.pdf index f1e5fddd7edee62110280ca65710a39b35344eb2..687c9d9cc3202ef8e46eb87173fe6be06dda2552 100644 --- a/data/processed/RQ3/Violin/landing_page_generator-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/landing_page_generator-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fff065fe232a66263554a640fa3bfb4e01989385b37f74f180ede28c24200a45 -size 35633 +oid sha256:6f998ba58a299ac49844c84fe663a988929e1b94428e426c49ebe41f8b6ec057 +size 35899 diff --git a/data/processed/RQ3/Violin/markdown_validator-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/markdown_validator-MCP_total_tokens_violin.pdf index 17d5489943ac1b317004bece568b2a460cf79494..7575c2be080d7f9b3473403f0c463b6d225a7767 100644 --- a/data/processed/RQ3/Violin/markdown_validator-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/markdown_validator-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:0a2df8a71ff9dc7b01a792376a1200cbce6109a099948ae863f88ff674211ae8 -size 30216 +oid sha256:7bead0f0de8074263cde20783bc25e5fa08bd17a1d5bb81d13fc1a38ed9e7c00 +size 30488 diff --git a/data/processed/RQ3/Violin/markdown_validator_total_tokens_violin.pdf b/data/processed/RQ3/Violin/markdown_validator_total_tokens_violin.pdf index 5ad0e5f57aa40fd63c49aad0c0fb7438a5098359..e51b4156e1ac4cebe5ff9d6f22518ea53d41341f 100644 --- a/data/processed/RQ3/Violin/markdown_validator_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/markdown_validator_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c16438d7456652c87061f7fc2c06f4b75e04096a5844f60dd705464e541ad2a5 -size 30371 +oid sha256:ad58e6a12fd6fa4f358df6a565f7e654e90eaa99baf737a663e8004c00a740a0 +size 30638 diff --git a/data/processed/RQ3/Violin/project_model_distribution.md b/data/processed/RQ3/Violin/project_model_distribution.md index 7f1d35390da22f3879635ad59c2b4a037c7512a5..51cbbe0db9bf459407f79a733d6177ab07d13d2e 100644 --- a/data/processed/RQ3/Violin/project_model_distribution.md +++ b/data/processed/RQ3/Violin/project_model_distribution.md @@ -2,258 +2,166 @@ Per-project, per-model total token usage with breakdown by execution outcome. Only the mean total tokens are reported. Baseline vs maximum is computed from the overall mean aggregated across all three statuses. -## SQL_assistant-A2A +## BookWriter-MCP ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 11,189 | -| GPT-4o-mini | 60 | 18,635 | -| DeepSeek-V3.1 | 60 | 36,805 | -| DeepSeek-R1 | 60 | 96,469 | -| Gemini-2.5 | 60 | 32,531 | -| Gemini-2.5-NT | 60 | 25,782 | -| Qwen3-235b | 60 | 17,381 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 36 | 8,736 | -| GPT-4o-mini | 50 | 17,092 | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 55,457 | -| Gemini-2.5 | 2 | 23,118 | -| Gemini-2.5-NT | 2 | 36,636 | -| Qwen3-235b | 35 | 13,848 | +| GPT-5 | 60 | 190,842 | +| GPT-4o-mini | 60 | 133,339 | +| DeepSeek-V3.1 | 60 | 184,315 | +| DeepSeek-R1 | 60 | 190,705 | +| Gemini-2.5 | 60 | 107,154 | +| Gemini-2.5-NT | 60 | 82,546 | +| Qwen3-235b | 60 | 110,716 | -### Pass (with retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 24 | 14,869 | -| GPT-4o-mini | 10 | 26,350 | -| DeepSeek-V3.1 | 60 | 36,805 | -| DeepSeek-R1 | 57 | 98,627 | -| Gemini-2.5 | 20 | 44,005 | -| Gemini-2.5-NT | 5 | 71,593 | -| Qwen3-235b | 23 | 23,436 | +| GPT-5 | 60 | 190,842 | +| GPT-4o-mini | 60 | 133,339 | +| DeepSeek-V3.1 | 60 | 184,315 | +| DeepSeek-R1 | 52 | 171,033 | +| Gemini-2.5 | 60 | 107,154 | +| Gemini-2.5-NT | 59 | 82,348 | +| Qwen3-235b | 59 | 110,871 | -### Failure distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 38 | 26,988 | -| Gemini-2.5-NT | 53 | 21,051 | -| Qwen3-235b | 2 | 9,587 | +| DeepSeek-R1 | 8 | 318,572 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 1 | 94,226 | +| Qwen3-235b | 1 | 101,570 | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 11,189 tokens -- Maximum (highest mean): DeepSeek-R1 = 96,469 tokens -- Delta: 85,280 (762.2%) +- Baseline (lowest mean): Gemini-2.5-NT = 82,546 tokens +- Maximum (highest mean): GPT-5 = 190,842 tokens +- Delta: 108,296 (131.2%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 8,736 tokens -- Maximum (highest mean): DeepSeek-R1 = 55,457 tokens -- Delta: 46,721 (534.8%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): Gemini-2.5-NT = 82,348 tokens +- Maximum (highest mean): GPT-5 = 190,842 tokens +- Delta: 108,494 (131.8%) -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): GPT-5 = 14,869 tokens -- Maximum (highest mean): DeepSeek-R1 = 98,627 tokens -- Delta: 83,759 (563.3%) +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): Gemini-2.5-NT = 94,226 tokens +- Maximum (highest mean): DeepSeek-R1 = 318,572 tokens +- Delta: 224,346 (238.1%) ### Baseline vs maximum — Failure -- Baseline (lowest mean): Qwen3-235b = 9,587 tokens -- Maximum (highest mean): Gemini-2.5 = 26,988 tokens -- Delta: 17,401 (181.5%) +- No data to compare. -## SQL_assistant-A2A_mix +## BookWriter-A2A ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 8,956 | -| GPT-4o-mini | 60 | 13,734 | -| DeepSeek-V3.1 | 60 | 16,479 | -| DeepSeek-R1 | 60 | 69,364 | -| Gemini-2.5 | 60 | 24,448 | -| Gemini-2.5-NT | 60 | 18,476 | -| Qwen3-235b | 60 | 14,779 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 46 | 6,953 | -| GPT-4o-mini | 43 | 10,004 | -| DeepSeek-V3.1 | 44 | 12,571 | -| DeepSeek-R1 | 20 | 29,145 | -| Gemini-2.5 | 25 | 14,332 | -| Gemini-2.5-NT | 33 | 9,561 | -| Qwen3-235b | 40 | 10,594 | +| GPT-5 | 60 | 181,022 | +| GPT-4o-mini | 60 | 139,665 | +| DeepSeek-V3.1 | 60 | 272,275 | +| DeepSeek-R1 | 60 | 186,678 | +| Gemini-2.5 | 60 | 110,126 | +| Gemini-2.5-NT | 60 | 86,249 | +| Qwen3-235b | 60 | 118,406 | -### Pass (with retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 14 | 15,537 | -| GPT-4o-mini | 17 | 23,168 | -| DeepSeek-V3.1 | 16 | 27,226 | -| DeepSeek-R1 | 40 | 89,473 | -| Gemini-2.5 | 26 | 30,481 | -| Gemini-2.5-NT | 19 | 33,544 | -| Qwen3-235b | 20 | 23,149 | +| GPT-5 | 60 | 181,022 | +| GPT-4o-mini | 60 | 139,665 | +| DeepSeek-V3.1 | 59 | 270,142 | +| DeepSeek-R1 | 52 | 173,528 | +| Gemini-2.5 | 60 | 110,126 | +| Gemini-2.5-NT | 60 | 86,249 | +| Qwen3-235b | 60 | 118,406 | -### Failure distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 9 | 35,120 | -| Gemini-2.5-NT | 8 | 19,462 | +| DeepSeek-V3.1 | 1 | 398,139 | +| DeepSeek-R1 | 8 | 272,152 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | | Qwen3-235b | 0 | - | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 8,956 tokens -- Maximum (highest mean): DeepSeek-R1 = 69,364 tokens -- Delta: 60,408 (674.5%) +- Baseline (lowest mean): Gemini-2.5-NT = 86,249 tokens +- Maximum (highest mean): DeepSeek-V3.1 = 272,275 tokens +- Delta: 186,026 (215.7%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 6,953 tokens -- Maximum (highest mean): DeepSeek-R1 = 29,145 tokens -- Delta: 22,193 (319.2%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): Gemini-2.5-NT = 86,249 tokens +- Maximum (highest mean): DeepSeek-V3.1 = 270,142 tokens +- Delta: 183,893 (213.2%) -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): GPT-5 = 15,537 tokens -- Maximum (highest mean): DeepSeek-R1 = 89,473 tokens -- Delta: 73,936 (475.9%) +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): DeepSeek-R1 = 272,152 tokens +- Maximum (highest mean): DeepSeek-V3.1 = 398,139 tokens +- Delta: 125,988 (46.3%) ### Baseline vs maximum — Failure -- Baseline (lowest mean): Gemini-2.5-NT = 19,462 tokens -- Maximum (highest mean): Gemini-2.5 = 35,120 tokens -- Delta: 15,658 (80.5%) +- No data to compare. -## SQL_assistant-MCP +## BookWriter-H-A2A ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 13,881 | -| GPT-4o-mini | 60 | 24,241 | -| DeepSeek-V3.1 | 60 | 25,637 | -| DeepSeek-R1 | 60 | 80,522 | -| Gemini-2.5 | 60 | 19,110 | -| Gemini-2.5-NT | 60 | 15,014 | -| Qwen3-235b | 60 | 17,232 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 19 | 6,378 | -| GPT-4o-mini | 21 | 14,334 | -| DeepSeek-V3.1 | 26 | 15,771 | -| DeepSeek-R1 | 13 | 35,319 | -| Gemini-2.5 | 6 | 15,321 | -| Gemini-2.5-NT | 10 | 12,100 | -| Qwen3-235b | 37 | 12,304 | +| GPT-5 | 60 | 117,358 | +| GPT-4o-mini | 60 | 102,564 | +| DeepSeek-V3.1 | 60 | 141,204 | +| DeepSeek-R1 | 60 | 129,883 | +| Gemini-2.5 | 60 | 91,585 | +| Gemini-2.5-NT | 60 | 65,826 | +| Qwen3-235b | 60 | 76,512 | -### Pass (with retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 41 | 17,358 | -| GPT-4o-mini | 39 | 29,576 | -| DeepSeek-V3.1 | 34 | 33,181 | -| DeepSeek-R1 | 47 | 93,025 | -| Gemini-2.5 | 11 | 42,818 | -| Gemini-2.5-NT | 6 | 33,419 | -| Qwen3-235b | 23 | 25,159 | +| GPT-5 | 60 | 117,358 | +| GPT-4o-mini | 60 | 102,564 | +| DeepSeek-V3.1 | 60 | 141,204 | +| DeepSeek-R1 | 59 | 128,948 | +| Gemini-2.5 | 60 | 91,585 | +| Gemini-2.5-NT | 60 | 65,826 | +| Qwen3-235b | 60 | 76,512 | -### Failure distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 43 | 13,573 | -| Gemini-2.5-NT | 44 | 13,166 | -| Qwen3-235b | 0 | - | - -### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 13,881 tokens -- Maximum (highest mean): DeepSeek-R1 = 80,522 tokens -- Delta: 66,641 (480.1%) - -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 6,378 tokens -- Maximum (highest mean): DeepSeek-R1 = 35,319 tokens -- Delta: 28,941 (453.8%) - -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): GPT-5 = 17,358 tokens -- Maximum (highest mean): DeepSeek-R1 = 93,025 tokens -- Delta: 75,667 (435.9%) - -### Baseline vs maximum — Failure -- Baseline (lowest mean): Gemini-2.5-NT = 13,166 tokens -- Maximum (highest mean): Gemini-2.5 = 13,573 tokens -- Delta: 407 (3.1%) - -## email_auto_responder_flow-MCP - -### Overall per-model total tokens (all statuses) -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 100 | 6,903 | -| GPT-4o-mini | 100 | 9,437 | -| DeepSeek-V3.1 | 100 | 9,580 | -| DeepSeek-R1 | 100 | 15,403 | -| Gemini-2.5 | 100 | 38,066 | -| Gemini-2.5-NT | 100 | 35,133 | -| Qwen3-235b | 100 | 7,434 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 100 | 6,903 | -| GPT-4o-mini | 100 | 9,437 | -| DeepSeek-V3.1 | 98 | 9,531 | -| DeepSeek-R1 | 96 | 14,798 | -| Gemini-2.5 | 91 | 33,629 | -| Gemini-2.5-NT | 88 | 12,696 | -| Qwen3-235b | 100 | 7,434 | - -### Pass (with retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 2 | 11,981 | -| DeepSeek-R1 | 4 | 29,922 | -| Gemini-2.5 | 9 | 82,929 | -| Gemini-2.5-NT | 12 | 199,673 | +| DeepSeek-R1 | 1 | 185,051 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | | Qwen3-235b | 0 | - | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 6,903 tokens -- Maximum (highest mean): Gemini-2.5 = 38,066 tokens -- Delta: 31,163 (451.4%) +- Baseline (lowest mean): Gemini-2.5-NT = 65,826 tokens +- Maximum (highest mean): DeepSeek-V3.1 = 141,204 tokens +- Delta: 75,378 (114.5%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 6,903 tokens -- Maximum (highest mean): Gemini-2.5 = 33,629 tokens -- Delta: 26,726 (387.2%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): Gemini-2.5-NT = 65,826 tokens +- Maximum (highest mean): DeepSeek-V3.1 = 141,204 tokens +- Delta: 75,378 (114.5%) -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): DeepSeek-V3.1 = 11,981 tokens -- Maximum (highest mean): Gemini-2.5-NT = 199,673 tokens -- Delta: 187,692 (1566.6%) +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): DeepSeek-R1 = 185,051 tokens +- Maximum (highest mean): DeepSeek-R1 = 185,051 tokens +- Delta: 0 (0.0%) ### Baseline vs maximum — Failure - No data to compare. -## email_auto_responder_flow +## EmailResponder ### Overall per-model total tokens (all statuses) | Model | n | Mean | @@ -266,7 +174,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 100 | 42,421 | | Qwen3-235b | 100 | 5,900 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 100 | 5,562 | @@ -277,7 +185,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 88 | 12,494 | | Qwen3-235b | 100 | 5,900 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | @@ -304,12 +212,12 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): Gemini-2.5-NT = 42,421 tokens - Delta: 36,859 (662.8%) -### Baseline vs maximum — Pass (no retries) +### Baseline vs maximum — Direct Success - Baseline (lowest mean): GPT-5 = 5,562 tokens - Maximum (highest mean): DeepSeek-R1 = 13,490 tokens - Delta: 7,929 (142.6%) -### Baseline vs maximum — Pass (with retries) +### Baseline vs maximum — Retry Success - Baseline (lowest mean): DeepSeek-V3.1 = 11,414 tokens - Maximum (highest mean): Gemini-2.5-NT = 261,883 tokens - Delta: 250,469 (2194.5%) @@ -319,47 +227,60 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): Gemini-2.5 = 405,135 tokens - Delta: 388,633 (2355.1%) -## game_builder-MCP +## EmailResponder-MCP ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 25 | 22,425 | -| GPT-4o-mini | 25 | 47,996 | -| DeepSeek-V3.1 | 25 | 55,289 | -| DeepSeek-R1 | 25 | 44,851 | -| Gemini-2.5 | 25 | 75,519 | -| Gemini-2.5-NT | 25 | 50,782 | -| Qwen3-235b | 25 | 24,039 | +| GPT-5 | 100 | 6,903 | +| GPT-4o-mini | 100 | 9,437 | +| DeepSeek-V3.1 | 100 | 9,580 | +| DeepSeek-R1 | 100 | 15,403 | +| Gemini-2.5 | 100 | 38,066 | +| Gemini-2.5-NT | 100 | 35,133 | +| Qwen3-235b | 100 | 7,434 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 25 | 22,425 | -| GPT-4o-mini | 25 | 47,996 | -| DeepSeek-V3.1 | 25 | 55,289 | -| DeepSeek-R1 | 25 | 44,851 | -| Gemini-2.5 | 25 | 75,519 | -| Gemini-2.5-NT | 25 | 50,782 | -| Qwen3-235b | 25 | 24,039 | +| GPT-5 | 100 | 6,903 | +| GPT-4o-mini | 100 | 9,437 | +| DeepSeek-V3.1 | 98 | 9,531 | +| DeepSeek-R1 | 96 | 14,798 | +| Gemini-2.5 | 91 | 33,629 | +| Gemini-2.5-NT | 88 | 12,696 | +| Qwen3-235b | 100 | 7,434 | + +### Retry Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 2 | 11,981 | +| DeepSeek-R1 | 4 | 29,922 | +| Gemini-2.5 | 9 | 82,929 | +| Gemini-2.5-NT | 12 | 199,673 | +| Qwen3-235b | 0 | - | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 22,425 tokens -- Maximum (highest mean): Gemini-2.5 = 75,519 tokens -- Delta: 53,093 (236.8%) +- Baseline (lowest mean): GPT-5 = 6,903 tokens +- Maximum (highest mean): Gemini-2.5 = 38,066 tokens +- Delta: 31,163 (451.4%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 22,425 tokens -- Maximum (highest mean): Gemini-2.5 = 75,519 tokens -- Delta: 53,093 (236.8%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 6,903 tokens +- Maximum (highest mean): Gemini-2.5 = 33,629 tokens +- Delta: 26,726 (387.2%) -### Baseline vs maximum — Pass (with retries) -- No data to compare. +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): DeepSeek-V3.1 = 11,981 tokens +- Maximum (highest mean): Gemini-2.5-NT = 199,673 tokens +- Delta: 187,692 (1566.6%) ### Baseline vs maximum — Failure - No data to compare. -## game_builder +## GameBuilder ### Overall per-model total tokens (all statuses) | Model | n | Mean | @@ -372,7 +293,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 25 | 67,964 | | Qwen3-235b | 25 | 24,212 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 25 | 22,904 | @@ -383,7 +304,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 23 | 54,423 | | Qwen3-235b | 25 | 24,212 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | @@ -410,12 +331,12 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): Gemini-2.5 = 72,772 tokens - Delta: 54,949 (308.3%) -### Baseline vs maximum — Pass (no retries) +### Baseline vs maximum — Direct Success - Baseline (lowest mean): GPT-4o-mini = 17,823 tokens - Maximum (highest mean): Gemini-2.5 = 72,772 tokens - Delta: 54,949 (308.3%) -### Baseline vs maximum — Pass (with retries) +### Baseline vs maximum — Retry Success - Baseline (lowest mean): Gemini-2.5-NT = 223,682 tokens - Maximum (highest mean): Gemini-2.5-NT = 223,682 tokens - Delta: 0 (0.0%) @@ -425,192 +346,113 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): DeepSeek-V3.1 = 77,354 tokens - Delta: 0 (0.0%) -## intelligent_recruitment_platform-A2A - -### Overall per-model total tokens (all statuses) -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 31,991 | -| GPT-4o-mini | 60 | 74,951 | -| DeepSeek-V3.1 | 60 | 65,662 | -| DeepSeek-R1 | 60 | 72,968 | -| Gemini-2.5 | 60 | 68,719 | -| Gemini-2.5-NT | 60 | 55,660 | -| Qwen3-235b | 60 | 97,717 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 31,991 | -| GPT-4o-mini | 60 | 74,951 | -| DeepSeek-V3.1 | 60 | 65,662 | -| DeepSeek-R1 | 56 | 73,964 | -| Gemini-2.5 | 60 | 68,719 | -| Gemini-2.5-NT | 60 | 55,660 | -| Qwen3-235b | 44 | 73,316 | - -### Pass (with retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 68,009 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 15 | 164,476 | - -### Failure distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 1 | 32,043 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 1 | 169,975 | - -### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 31,991 tokens -- Maximum (highest mean): Qwen3-235b = 97,717 tokens -- Delta: 65,726 (205.5%) - -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 31,991 tokens -- Maximum (highest mean): GPT-4o-mini = 74,951 tokens -- Delta: 42,960 (134.3%) - -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): DeepSeek-R1 = 68,009 tokens -- Maximum (highest mean): Qwen3-235b = 164,476 tokens -- Delta: 96,467 (141.8%) - -### Baseline vs maximum — Failure -- Baseline (lowest mean): DeepSeek-R1 = 32,043 tokens -- Maximum (highest mean): Qwen3-235b = 169,975 tokens -- Delta: 137,932 (430.5%) - -## intelligent_recruitment_platform-A2A_mix +## GameBuilder-MCP ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 32,672 | -| GPT-4o-mini | 60 | 101,936 | -| DeepSeek-V3.1 | 60 | 65,048 | -| DeepSeek-R1 | 60 | 79,028 | -| Gemini-2.5 | 60 | 57,128 | -| Gemini-2.5-NT | 60 | 36,887 | -| Qwen3-235b | 60 | 73,259 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 32,672 | -| GPT-4o-mini | 60 | 101,936 | -| DeepSeek-V3.1 | 60 | 65,048 | -| DeepSeek-R1 | 57 | 76,741 | -| Gemini-2.5 | 60 | 57,128 | -| Gemini-2.5-NT | 60 | 36,887 | -| Qwen3-235b | 49 | 67,089 | +| GPT-5 | 25 | 22,425 | +| GPT-4o-mini | 25 | 47,996 | +| DeepSeek-V3.1 | 25 | 55,289 | +| DeepSeek-R1 | 25 | 44,851 | +| Gemini-2.5 | 25 | 75,519 | +| Gemini-2.5-NT | 25 | 50,782 | +| Qwen3-235b | 25 | 24,039 | -### Pass (with retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 122,480 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 11 | 100,743 | +| GPT-5 | 25 | 22,425 | +| GPT-4o-mini | 25 | 47,996 | +| DeepSeek-V3.1 | 25 | 55,289 | +| DeepSeek-R1 | 25 | 44,851 | +| Gemini-2.5 | 25 | 75,519 | +| Gemini-2.5-NT | 25 | 50,782 | +| Qwen3-235b | 25 | 24,039 | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 32,672 tokens -- Maximum (highest mean): GPT-4o-mini = 101,936 tokens -- Delta: 69,264 (212.0%) +- Baseline (lowest mean): GPT-5 = 22,425 tokens +- Maximum (highest mean): Gemini-2.5 = 75,519 tokens +- Delta: 53,093 (236.8%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 32,672 tokens -- Maximum (highest mean): GPT-4o-mini = 101,936 tokens -- Delta: 69,264 (212.0%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 22,425 tokens +- Maximum (highest mean): Gemini-2.5 = 75,519 tokens +- Delta: 53,093 (236.8%) -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): Qwen3-235b = 100,743 tokens -- Maximum (highest mean): DeepSeek-R1 = 122,480 tokens -- Delta: 21,738 (21.6%) +### Baseline vs maximum — Retry Success +- No data to compare. ### Baseline vs maximum — Failure - No data to compare. -## intelligent_recruitment_platform-MCP +## LandingPageGenerator-MCP ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 32,411 | -| GPT-4o-mini | 60 | 99,933 | -| DeepSeek-V3.1 | 60 | 65,748 | -| DeepSeek-R1 | 60 | 84,686 | -| Gemini-2.5 | 60 | 64,864 | -| Gemini-2.5-NT | 60 | 54,402 | -| Qwen3-235b | 60 | 87,098 | +| GPT-5 | 60 | 12,653 | +| GPT-4o-mini | 60 | 19,250 | +| DeepSeek-V3.1 | 60 | 15,866 | +| DeepSeek-R1 | 60 | 28,371 | +| Gemini-2.5 | 60 | 26,527 | +| Gemini-2.5-NT | 60 | 33,428 | +| Qwen3-235b | 60 | 14,478 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 32,411 | -| GPT-4o-mini | 60 | 99,933 | -| DeepSeek-V3.1 | 60 | 65,748 | -| DeepSeek-R1 | 58 | 82,498 | -| Gemini-2.5 | 60 | 64,864 | -| Gemini-2.5-NT | 60 | 54,402 | -| Qwen3-235b | 40 | 68,264 | +| GPT-5 | 60 | 12,653 | +| GPT-4o-mini | 59 | 19,333 | +| DeepSeek-V3.1 | 55 | 15,912 | +| DeepSeek-R1 | 33 | 24,081 | +| Gemini-2.5 | 1 | 55,158 | +| Gemini-2.5-NT | 2 | 188,992 | +| Qwen3-235b | 53 | 15,068 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 2 | 148,137 | +| DeepSeek-R1 | 8 | 61,223 | | Gemini-2.5 | 0 | - | | Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 18 | 124,320 | +| Qwen3-235b | 0 | - | ### Failure distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 2 | 128,765 | +| GPT-4o-mini | 1 | 14,308 | +| DeepSeek-V3.1 | 5 | 15,362 | +| DeepSeek-R1 | 19 | 21,990 | +| Gemini-2.5 | 59 | 26,041 | +| Gemini-2.5-NT | 58 | 28,064 | +| Qwen3-235b | 7 | 10,009 | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 32,411 tokens -- Maximum (highest mean): GPT-4o-mini = 99,933 tokens -- Delta: 67,523 (208.3%) +- Baseline (lowest mean): GPT-5 = 12,653 tokens +- Maximum (highest mean): Gemini-2.5-NT = 33,428 tokens +- Delta: 20,775 (164.2%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 32,411 tokens -- Maximum (highest mean): GPT-4o-mini = 99,933 tokens -- Delta: 67,523 (208.3%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 12,653 tokens +- Maximum (highest mean): Gemini-2.5-NT = 188,992 tokens +- Delta: 176,339 (1393.6%) -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): Qwen3-235b = 124,320 tokens -- Maximum (highest mean): DeepSeek-R1 = 148,137 tokens -- Delta: 23,817 (19.2%) +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): DeepSeek-R1 = 61,223 tokens +- Maximum (highest mean): DeepSeek-R1 = 61,223 tokens +- Delta: 0 (0.0%) ### Baseline vs maximum — Failure -- Baseline (lowest mean): Qwen3-235b = 128,765 tokens -- Maximum (highest mean): Qwen3-235b = 128,765 tokens -- Delta: 0 (0.0%) +- Baseline (lowest mean): Qwen3-235b = 10,009 tokens +- Maximum (highest mean): Gemini-2.5-NT = 28,064 tokens +- Delta: 18,055 (180.4%) -## landing_page_generator-A2A +## LandingPageGenerator-A2A ### Overall per-model total tokens (all statuses) | Model | n | Mean | @@ -623,7 +465,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 60 | 22,826 | | Qwen3-235b | 60 | 18,142 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 16 | 18,863 | @@ -634,7 +476,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 27 | 20,066 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | @@ -661,12 +503,12 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): DeepSeek-V3.1 = 28,814 tokens - Delta: 13,849 (92.5%) -### Baseline vs maximum — Pass (no retries) +### Baseline vs maximum — Direct Success - Baseline (lowest mean): GPT-5 = 18,863 tokens - Maximum (highest mean): DeepSeek-R1 = 46,577 tokens - Delta: 27,714 (146.9%) -### Baseline vs maximum — Pass (with retries) +### Baseline vs maximum — Retry Success - Baseline (lowest mean): Qwen3-235b = 34,203 tokens - Maximum (highest mean): Qwen3-235b = 34,203 tokens - Delta: 0 (0.0%) @@ -676,7 +518,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): GPT-4o-mini = 27,976 tokens - Delta: 14,429 (106.5%) -## landing_page_generator-A2A_mix +## LandingPageGenerator-H-A2A ### Overall per-model total tokens (all statuses) | Model | n | Mean | @@ -689,7 +531,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 60 | 14,949 | | Qwen3-235b | 60 | 17,628 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 25 | 29,492 | @@ -700,7 +542,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 29 | 19,428 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | @@ -727,12 +569,12 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): DeepSeek-R1 = 57,203 tokens - Delta: 42,254 (282.6%) -### Baseline vs maximum — Pass (no retries) +### Baseline vs maximum — Direct Success - Baseline (lowest mean): Qwen3-235b = 19,428 tokens - Maximum (highest mean): DeepSeek-R1 = 68,378 tokens - Delta: 48,950 (252.0%) -### Baseline vs maximum — Pass (with retries) +### Baseline vs maximum — Retry Success - Baseline (lowest mean): DeepSeek-V3.1 = 28,430 tokens - Maximum (highest mean): Qwen3-235b = 33,284 tokens - Delta: 4,854 (17.1%) @@ -742,73 +584,60 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): DeepSeek-R1 = 55,231 tokens - Delta: 40,483 (274.5%) -## landing_page_generator-MCP +## MarkdownValidator ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 12,653 | -| GPT-4o-mini | 60 | 19,250 | -| DeepSeek-V3.1 | 60 | 15,866 | -| DeepSeek-R1 | 60 | 28,371 | -| Gemini-2.5 | 60 | 26,527 | -| Gemini-2.5-NT | 60 | 33,428 | -| Qwen3-235b | 60 | 14,478 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 12,653 | -| GPT-4o-mini | 59 | 19,333 | -| DeepSeek-V3.1 | 55 | 15,912 | -| DeepSeek-R1 | 33 | 24,081 | -| Gemini-2.5 | 1 | 55,158 | -| Gemini-2.5-NT | 2 | 188,992 | -| Qwen3-235b | 53 | 15,068 | +| GPT-5 | 60 | 1,853 | +| GPT-4o-mini | 60 | 1,853 | +| DeepSeek-V3.1 | 60 | 1,948 | +| DeepSeek-R1 | 60 | 5,912 | +| Gemini-2.5 | 60 | 135,224 | +| Gemini-2.5-NT | 60 | 31,330 | +| Qwen3-235b | 60 | 1,854 | -### Pass (with retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 8 | 61,223 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 0 | - | +| GPT-5 | 60 | 1,853 | +| GPT-4o-mini | 60 | 1,853 | +| DeepSeek-V3.1 | 60 | 1,948 | +| DeepSeek-R1 | 60 | 5,912 | +| Gemini-2.5 | 47 | 81,437 | +| Gemini-2.5-NT | 57 | 16,246 | +| Qwen3-235b | 60 | 1,854 | -### Failure distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | -| GPT-4o-mini | 1 | 14,308 | -| DeepSeek-V3.1 | 5 | 15,362 | -| DeepSeek-R1 | 19 | 21,990 | -| Gemini-2.5 | 59 | 26,041 | -| Gemini-2.5-NT | 58 | 28,064 | -| Qwen3-235b | 7 | 10,009 | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 13 | 329,685 | +| Gemini-2.5-NT | 3 | 317,934 | +| Qwen3-235b | 0 | - | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 12,653 tokens -- Maximum (highest mean): Gemini-2.5-NT = 33,428 tokens -- Delta: 20,775 (164.2%) +- Baseline (lowest mean): GPT-5 = 1,853 tokens +- Maximum (highest mean): Gemini-2.5 = 135,224 tokens +- Delta: 133,372 (7199.5%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 12,653 tokens -- Maximum (highest mean): Gemini-2.5-NT = 188,992 tokens -- Delta: 176,339 (1393.6%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 1,853 tokens +- Maximum (highest mean): Gemini-2.5 = 81,437 tokens +- Delta: 79,585 (4296.0%) -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): DeepSeek-R1 = 61,223 tokens -- Maximum (highest mean): DeepSeek-R1 = 61,223 tokens -- Delta: 0 (0.0%) +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): Gemini-2.5-NT = 317,934 tokens +- Maximum (highest mean): Gemini-2.5 = 329,685 tokens +- Delta: 11,751 (3.7%) ### Baseline vs maximum — Failure -- Baseline (lowest mean): Qwen3-235b = 10,009 tokens -- Maximum (highest mean): Gemini-2.5-NT = 28,064 tokens -- Delta: 18,055 (180.4%) +- No data to compare. -## markdown_validator-MCP +## MarkdownValidator-MCP ### Overall per-model total tokens (all statuses) | Model | n | Mean | @@ -821,7 +650,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 60 | 61,899 | | Qwen3-235b | 60 | 1,994 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 60 | 1,967 | @@ -832,7 +661,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 56 | 39,448 | | Qwen3-235b | 60 | 1,994 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | @@ -848,12 +677,12 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): Gemini-2.5 = 95,503 tokens - Delta: 93,561 (4819.0%) -### Baseline vs maximum — Pass (no retries) +### Baseline vs maximum — Direct Success - Baseline (lowest mean): GPT-4o-mini = 1,942 tokens - Maximum (highest mean): Gemini-2.5 = 67,759 tokens - Delta: 65,817 (3390.0%) -### Baseline vs maximum — Pass (with retries) +### Baseline vs maximum — Retry Success - Baseline (lowest mean): Gemini-2.5 = 206,480 tokens - Maximum (highest mean): Gemini-2.5-NT = 376,206 tokens - Delta: 169,726 (82.2%) @@ -861,60 +690,456 @@ Per-project, per-model total token usage with breakdown by execution outcome. On ### Baseline vs maximum — Failure - No data to compare. -## markdown_validator +## RecruitmentAssistant-MCP ### Overall per-model total tokens (all statuses) | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 1,853 | -| GPT-4o-mini | 60 | 1,853 | -| DeepSeek-V3.1 | 60 | 1,948 | -| DeepSeek-R1 | 60 | 5,912 | -| Gemini-2.5 | 60 | 135,224 | -| Gemini-2.5-NT | 60 | 31,330 | -| Qwen3-235b | 60 | 1,854 | +| GPT-5 | 60 | 32,411 | +| GPT-4o-mini | 60 | 99,933 | +| DeepSeek-V3.1 | 60 | 65,748 | +| DeepSeek-R1 | 60 | 84,686 | +| Gemini-2.5 | 60 | 64,864 | +| Gemini-2.5-NT | 60 | 54,402 | +| Qwen3-235b | 60 | 87,098 | + +### Direct Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 32,411 | +| GPT-4o-mini | 60 | 99,933 | +| DeepSeek-V3.1 | 60 | 65,748 | +| DeepSeek-R1 | 58 | 82,498 | +| Gemini-2.5 | 60 | 64,864 | +| Gemini-2.5-NT | 60 | 54,402 | +| Qwen3-235b | 40 | 68,264 | + +### Retry Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 2 | 148,137 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 18 | 124,320 | + +### Failure distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 2 | 128,765 | + +### Baseline vs maximum (overall mean tokens) +- Baseline (lowest mean): GPT-5 = 32,411 tokens +- Maximum (highest mean): GPT-4o-mini = 99,933 tokens +- Delta: 67,523 (208.3%) + +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 32,411 tokens +- Maximum (highest mean): GPT-4o-mini = 99,933 tokens +- Delta: 67,523 (208.3%) + +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): Qwen3-235b = 124,320 tokens +- Maximum (highest mean): DeepSeek-R1 = 148,137 tokens +- Delta: 23,817 (19.2%) + +### Baseline vs maximum — Failure +- Baseline (lowest mean): Qwen3-235b = 128,765 tokens +- Maximum (highest mean): Qwen3-235b = 128,765 tokens +- Delta: 0 (0.0%) + +## RecruitmentAssistant-A2A + +### Overall per-model total tokens (all statuses) +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 31,991 | +| GPT-4o-mini | 60 | 74,951 | +| DeepSeek-V3.1 | 60 | 65,662 | +| DeepSeek-R1 | 60 | 72,968 | +| Gemini-2.5 | 60 | 68,719 | +| Gemini-2.5-NT | 60 | 55,660 | +| Qwen3-235b | 60 | 97,717 | + +### Direct Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 31,991 | +| GPT-4o-mini | 60 | 74,951 | +| DeepSeek-V3.1 | 60 | 65,662 | +| DeepSeek-R1 | 56 | 73,964 | +| Gemini-2.5 | 60 | 68,719 | +| Gemini-2.5-NT | 60 | 55,660 | +| Qwen3-235b | 44 | 73,316 | + +### Retry Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 3 | 68,009 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 15 | 164,476 | + +### Failure distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 1 | 32,043 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 1 | 169,975 | + +### Baseline vs maximum (overall mean tokens) +- Baseline (lowest mean): GPT-5 = 31,991 tokens +- Maximum (highest mean): Qwen3-235b = 97,717 tokens +- Delta: 65,726 (205.5%) + +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 31,991 tokens +- Maximum (highest mean): GPT-4o-mini = 74,951 tokens +- Delta: 42,960 (134.3%) + +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): DeepSeek-R1 = 68,009 tokens +- Maximum (highest mean): Qwen3-235b = 164,476 tokens +- Delta: 96,467 (141.8%) + +### Baseline vs maximum — Failure +- Baseline (lowest mean): DeepSeek-R1 = 32,043 tokens +- Maximum (highest mean): Qwen3-235b = 169,975 tokens +- Delta: 137,932 (430.5%) + +## RecruitmentAssistant-H-A2A + +### Overall per-model total tokens (all statuses) +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 32,672 | +| GPT-4o-mini | 60 | 101,936 | +| DeepSeek-V3.1 | 60 | 65,048 | +| DeepSeek-R1 | 60 | 79,028 | +| Gemini-2.5 | 60 | 57,128 | +| Gemini-2.5-NT | 60 | 36,887 | +| Qwen3-235b | 60 | 73,259 | + +### Direct Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 32,672 | +| GPT-4o-mini | 60 | 101,936 | +| DeepSeek-V3.1 | 60 | 65,048 | +| DeepSeek-R1 | 57 | 76,741 | +| Gemini-2.5 | 60 | 57,128 | +| Gemini-2.5-NT | 60 | 36,887 | +| Qwen3-235b | 49 | 67,089 | + +### Retry Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 3 | 122,480 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 11 | 100,743 | + +### Baseline vs maximum (overall mean tokens) +- Baseline (lowest mean): GPT-5 = 32,672 tokens +- Maximum (highest mean): GPT-4o-mini = 101,936 tokens +- Delta: 69,264 (212.0%) + +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 32,672 tokens +- Maximum (highest mean): GPT-4o-mini = 101,936 tokens +- Delta: 69,264 (212.0%) + +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): Qwen3-235b = 100,743 tokens +- Maximum (highest mean): DeepSeek-R1 = 122,480 tokens +- Delta: 21,738 (21.6%) + +### Baseline vs maximum — Failure +- No data to compare. + +## SQLAssistant-MCP + +### Overall per-model total tokens (all statuses) +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 13,881 | +| GPT-4o-mini | 60 | 24,241 | +| DeepSeek-V3.1 | 60 | 25,637 | +| DeepSeek-R1 | 60 | 80,522 | +| Gemini-2.5 | 60 | 19,110 | +| Gemini-2.5-NT | 60 | 15,014 | +| Qwen3-235b | 60 | 17,232 | + +### Direct Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 19 | 6,378 | +| GPT-4o-mini | 21 | 14,334 | +| DeepSeek-V3.1 | 26 | 15,771 | +| DeepSeek-R1 | 13 | 35,319 | +| Gemini-2.5 | 6 | 15,321 | +| Gemini-2.5-NT | 10 | 12,100 | +| Qwen3-235b | 37 | 12,304 | + +### Retry Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 41 | 17,358 | +| GPT-4o-mini | 39 | 29,576 | +| DeepSeek-V3.1 | 34 | 33,181 | +| DeepSeek-R1 | 47 | 93,025 | +| Gemini-2.5 | 11 | 42,818 | +| Gemini-2.5-NT | 6 | 33,419 | +| Qwen3-235b | 23 | 25,159 | + +### Failure distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 43 | 13,573 | +| Gemini-2.5-NT | 44 | 13,166 | +| Qwen3-235b | 0 | - | + +### Baseline vs maximum (overall mean tokens) +- Baseline (lowest mean): GPT-5 = 13,881 tokens +- Maximum (highest mean): DeepSeek-R1 = 80,522 tokens +- Delta: 66,641 (480.1%) + +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 6,378 tokens +- Maximum (highest mean): DeepSeek-R1 = 35,319 tokens +- Delta: 28,941 (453.8%) + +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): GPT-5 = 17,358 tokens +- Maximum (highest mean): DeepSeek-R1 = 93,025 tokens +- Delta: 75,667 (435.9%) + +### Baseline vs maximum — Failure +- Baseline (lowest mean): Gemini-2.5-NT = 13,166 tokens +- Maximum (highest mean): Gemini-2.5 = 13,573 tokens +- Delta: 407 (3.1%) + +## SQLAssistant-A2A + +### Overall per-model total tokens (all statuses) +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 11,189 | +| GPT-4o-mini | 60 | 18,635 | +| DeepSeek-V3.1 | 60 | 36,805 | +| DeepSeek-R1 | 60 | 96,469 | +| Gemini-2.5 | 60 | 32,531 | +| Gemini-2.5-NT | 60 | 25,782 | +| Qwen3-235b | 60 | 17,381 | + +### Direct Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 36 | 8,736 | +| GPT-4o-mini | 50 | 17,092 | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 3 | 55,457 | +| Gemini-2.5 | 2 | 23,118 | +| Gemini-2.5-NT | 2 | 36,636 | +| Qwen3-235b | 35 | 13,848 | + +### Retry Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 24 | 14,869 | +| GPT-4o-mini | 10 | 26,350 | +| DeepSeek-V3.1 | 60 | 36,805 | +| DeepSeek-R1 | 57 | 98,627 | +| Gemini-2.5 | 20 | 44,005 | +| Gemini-2.5-NT | 5 | 71,593 | +| Qwen3-235b | 23 | 23,436 | + +### Failure distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 38 | 26,988 | +| Gemini-2.5-NT | 53 | 21,051 | +| Qwen3-235b | 2 | 9,587 | + +### Baseline vs maximum (overall mean tokens) +- Baseline (lowest mean): GPT-5 = 11,189 tokens +- Maximum (highest mean): DeepSeek-R1 = 96,469 tokens +- Delta: 85,280 (762.2%) + +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 8,736 tokens +- Maximum (highest mean): DeepSeek-R1 = 55,457 tokens +- Delta: 46,721 (534.8%) + +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): GPT-5 = 14,869 tokens +- Maximum (highest mean): DeepSeek-R1 = 98,627 tokens +- Delta: 83,759 (563.3%) + +### Baseline vs maximum — Failure +- Baseline (lowest mean): Qwen3-235b = 9,587 tokens +- Maximum (highest mean): Gemini-2.5 = 26,988 tokens +- Delta: 17,401 (181.5%) + +## SQLAssistant-H-A2A + +### Overall per-model total tokens (all statuses) +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 8,956 | +| GPT-4o-mini | 60 | 13,734 | +| DeepSeek-V3.1 | 60 | 16,479 | +| DeepSeek-R1 | 60 | 69,364 | +| Gemini-2.5 | 60 | 24,448 | +| Gemini-2.5-NT | 60 | 18,476 | +| Qwen3-235b | 60 | 14,779 | + +### Direct Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 46 | 6,953 | +| GPT-4o-mini | 43 | 10,004 | +| DeepSeek-V3.1 | 44 | 12,571 | +| DeepSeek-R1 | 20 | 29,145 | +| Gemini-2.5 | 25 | 14,332 | +| Gemini-2.5-NT | 33 | 9,561 | +| Qwen3-235b | 40 | 10,594 | + +### Retry Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 14 | 15,537 | +| GPT-4o-mini | 17 | 23,168 | +| DeepSeek-V3.1 | 16 | 27,226 | +| DeepSeek-R1 | 40 | 89,473 | +| Gemini-2.5 | 26 | 30,481 | +| Gemini-2.5-NT | 19 | 33,544 | +| Qwen3-235b | 20 | 23,149 | + +### Failure distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 9 | 35,120 | +| Gemini-2.5-NT | 8 | 19,462 | +| Qwen3-235b | 0 | - | + +### Baseline vs maximum (overall mean tokens) +- Baseline (lowest mean): GPT-5 = 8,956 tokens +- Maximum (highest mean): DeepSeek-R1 = 69,364 tokens +- Delta: 60,408 (674.5%) + +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): GPT-5 = 6,953 tokens +- Maximum (highest mean): DeepSeek-R1 = 29,145 tokens +- Delta: 22,193 (319.2%) + +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): GPT-5 = 15,537 tokens +- Maximum (highest mean): DeepSeek-R1 = 89,473 tokens +- Delta: 73,936 (475.9%) + +### Baseline vs maximum — Failure +- Baseline (lowest mean): Gemini-2.5-NT = 19,462 tokens +- Maximum (highest mean): Gemini-2.5 = 35,120 tokens +- Delta: 15,658 (80.5%) + +## SocialMediaManager-MCP + +### Overall per-model total tokens (all statuses) +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 60 | 38,940 | +| GPT-4o-mini | 60 | 108,089 | +| DeepSeek-V3.1 | 60 | 95,721 | +| DeepSeek-R1 | 60 | 109,146 | +| Gemini-2.5 | 60 | 12,749 | +| Gemini-2.5-NT | 60 | 11,480 | +| Qwen3-235b | 60 | 46,557 | + +### Direct Success distribution +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 33 | 22,287 | +| GPT-4o-mini | 20 | 40,555 | +| DeepSeek-V3.1 | 37 | 64,742 | +| DeepSeek-R1 | 21 | 41,849 | +| Gemini-2.5 | 51 | 11,390 | +| Gemini-2.5-NT | 55 | 10,904 | +| Qwen3-235b | 35 | 27,216 | -### Pass (no retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 1,853 | -| GPT-4o-mini | 60 | 1,853 | -| DeepSeek-V3.1 | 60 | 1,948 | -| DeepSeek-R1 | 60 | 5,912 | -| Gemini-2.5 | 47 | 81,437 | -| Gemini-2.5-NT | 57 | 16,246 | -| Qwen3-235b | 60 | 1,854 | +| GPT-5 | 26 | 57,749 | +| GPT-4o-mini | 37 | 134,943 | +| DeepSeek-V3.1 | 23 | 145,557 | +| DeepSeek-R1 | 36 | 138,306 | +| Gemini-2.5 | 9 | 20,447 | +| Gemini-2.5-NT | 5 | 17,821 | +| Qwen3-235b | 24 | 72,110 | -### Pass (with retries) distribution +### Failure distribution | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | +| GPT-5 | 1 | 99,433 | +| GPT-4o-mini | 3 | 227,117 | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 13 | 329,685 | -| Gemini-2.5-NT | 3 | 317,934 | -| Qwen3-235b | 0 | - | +| DeepSeek-R1 | 3 | 230,298 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 1 | 110,239 | ### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): GPT-5 = 1,853 tokens -- Maximum (highest mean): Gemini-2.5 = 135,224 tokens -- Delta: 133,372 (7199.5%) +- Baseline (lowest mean): Gemini-2.5-NT = 11,480 tokens +- Maximum (highest mean): DeepSeek-R1 = 109,146 tokens +- Delta: 97,666 (850.7%) -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): GPT-5 = 1,853 tokens -- Maximum (highest mean): Gemini-2.5 = 81,437 tokens -- Delta: 79,585 (4296.0%) +### Baseline vs maximum — Direct Success +- Baseline (lowest mean): Gemini-2.5-NT = 10,904 tokens +- Maximum (highest mean): DeepSeek-V3.1 = 64,742 tokens +- Delta: 53,839 (493.8%) -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): Gemini-2.5-NT = 317,934 tokens -- Maximum (highest mean): Gemini-2.5 = 329,685 tokens -- Delta: 11,751 (3.7%) +### Baseline vs maximum — Retry Success +- Baseline (lowest mean): Gemini-2.5-NT = 17,821 tokens +- Maximum (highest mean): DeepSeek-V3.1 = 145,557 tokens +- Delta: 127,736 (716.8%) ### Baseline vs maximum — Failure -- No data to compare. +- Baseline (lowest mean): GPT-5 = 99,433 tokens +- Maximum (highest mean): DeepSeek-R1 = 230,298 tokens +- Delta: 130,865 (131.6%) -## self_evaluation_loop_flow-A2A +## SocialMediaManager-A2A ### Overall per-model total tokens (all statuses) | Model | n | Mean | @@ -927,7 +1152,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 60 | 11,429 | | Qwen3-235b | 60 | 55,866 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 31 | 23,996 | @@ -938,7 +1163,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 53 | 10,675 | | Qwen3-235b | 27 | 30,147 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 27 | 55,285 | @@ -965,12 +1190,12 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): DeepSeek-R1 = 124,960 tokens - Delta: 113,531 (993.3%) -### Baseline vs maximum — Pass (no retries) +### Baseline vs maximum — Direct Success - Baseline (lowest mean): Gemini-2.5-NT = 10,675 tokens - Maximum (highest mean): DeepSeek-V3.1 = 70,407 tokens - Delta: 59,732 (559.5%) -### Baseline vs maximum — Pass (with retries) +### Baseline vs maximum — Retry Success - Baseline (lowest mean): Gemini-2.5-NT = 17,136 tokens - Maximum (highest mean): DeepSeek-R1 = 165,493 tokens - Delta: 148,357 (865.8%) @@ -980,7 +1205,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): Qwen3-235b = 230,991 tokens - Delta: 132,183 (133.8%) -## self_evaluation_loop_flow-A2A_mix +## SocialMediaManager-H-A2A ### Overall per-model total tokens (all statuses) | Model | n | Mean | @@ -993,7 +1218,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 60 | 9,642 | | Qwen3-235b | 60 | 29,440 | -### Pass (no retries) distribution +### Direct Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 55 | 15,857 | @@ -1004,7 +1229,7 @@ Per-project, per-model total token usage with breakdown by execution outcome. On | Gemini-2.5-NT | 45 | 7,487 | | Qwen3-235b | 35 | 14,019 | -### Pass (with retries) distribution +### Retry Success distribution | Model | n | Mean | | --- | --- | --- | | GPT-5 | 5 | 27,133 | @@ -1031,12 +1256,12 @@ Per-project, per-model total token usage with breakdown by execution outcome. On - Maximum (highest mean): DeepSeek-R1 = 98,225 tokens - Delta: 88,583 (918.7%) -### Baseline vs maximum — Pass (no retries) +### Baseline vs maximum — Direct Success - Baseline (lowest mean): Gemini-2.5-NT = 7,487 tokens - Maximum (highest mean): DeepSeek-V3.1 = 51,074 tokens - Delta: 43,587 (582.2%) -### Baseline vs maximum — Pass (with retries) +### Baseline vs maximum — Retry Success - Baseline (lowest mean): Gemini-2.5-NT = 16,107 tokens - Maximum (highest mean): DeepSeek-R1 = 133,368 tokens - Delta: 117,261 (728.0%) @@ -1044,229 +1269,4 @@ Per-project, per-model total token usage with breakdown by execution outcome. On ### Baseline vs maximum — Failure - Baseline (lowest mean): Qwen3-235b = 41,361 tokens - Maximum (highest mean): DeepSeek-R1 = 336,866 tokens -- Delta: 295,505 (714.5%) - -## self_evaluation_loop_flow-MCP - -### Overall per-model total tokens (all statuses) -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 38,940 | -| GPT-4o-mini | 60 | 108,089 | -| DeepSeek-V3.1 | 60 | 95,721 | -| DeepSeek-R1 | 60 | 109,146 | -| Gemini-2.5 | 60 | 12,749 | -| Gemini-2.5-NT | 60 | 11,480 | -| Qwen3-235b | 60 | 46,557 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 33 | 22,287 | -| GPT-4o-mini | 20 | 40,555 | -| DeepSeek-V3.1 | 37 | 64,742 | -| DeepSeek-R1 | 21 | 41,849 | -| Gemini-2.5 | 51 | 11,390 | -| Gemini-2.5-NT | 55 | 10,904 | -| Qwen3-235b | 35 | 27,216 | - -### Pass (with retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 26 | 57,749 | -| GPT-4o-mini | 37 | 134,943 | -| DeepSeek-V3.1 | 23 | 145,557 | -| DeepSeek-R1 | 36 | 138,306 | -| Gemini-2.5 | 9 | 20,447 | -| Gemini-2.5-NT | 5 | 17,821 | -| Qwen3-235b | 24 | 72,110 | - -### Failure distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 1 | 99,433 | -| GPT-4o-mini | 3 | 227,117 | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 230,298 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 1 | 110,239 | - -### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): Gemini-2.5-NT = 11,480 tokens -- Maximum (highest mean): DeepSeek-R1 = 109,146 tokens -- Delta: 97,666 (850.7%) - -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): Gemini-2.5-NT = 10,904 tokens -- Maximum (highest mean): DeepSeek-V3.1 = 64,742 tokens -- Delta: 53,839 (493.8%) - -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): Gemini-2.5-NT = 17,821 tokens -- Maximum (highest mean): DeepSeek-V3.1 = 145,557 tokens -- Delta: 127,736 (716.8%) - -### Baseline vs maximum — Failure -- Baseline (lowest mean): GPT-5 = 99,433 tokens -- Maximum (highest mean): DeepSeek-R1 = 230,298 tokens -- Delta: 130,865 (131.6%) - -## write_a_book_with_flows-A2A - -### Overall per-model total tokens (all statuses) -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 181,022 | -| GPT-4o-mini | 60 | 139,665 | -| DeepSeek-V3.1 | 60 | 272,275 | -| DeepSeek-R1 | 60 | 186,678 | -| Gemini-2.5 | 60 | 110,126 | -| Gemini-2.5-NT | 60 | 86,249 | -| Qwen3-235b | 60 | 118,406 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 181,022 | -| GPT-4o-mini | 60 | 139,665 | -| DeepSeek-V3.1 | 59 | 270,142 | -| DeepSeek-R1 | 52 | 173,528 | -| Gemini-2.5 | 60 | 110,126 | -| Gemini-2.5-NT | 60 | 86,249 | -| Qwen3-235b | 60 | 118,406 | - -### Pass (with retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 1 | 398,139 | -| DeepSeek-R1 | 8 | 272,152 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 0 | - | - -### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): Gemini-2.5-NT = 86,249 tokens -- Maximum (highest mean): DeepSeek-V3.1 = 272,275 tokens -- Delta: 186,026 (215.7%) - -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): Gemini-2.5-NT = 86,249 tokens -- Maximum (highest mean): DeepSeek-V3.1 = 270,142 tokens -- Delta: 183,893 (213.2%) - -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): DeepSeek-R1 = 272,152 tokens -- Maximum (highest mean): DeepSeek-V3.1 = 398,139 tokens -- Delta: 125,988 (46.3%) - -### Baseline vs maximum — Failure -- No data to compare. - -## write_a_book_with_flows-A2A_mix - -### Overall per-model total tokens (all statuses) -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 117,358 | -| GPT-4o-mini | 60 | 102,564 | -| DeepSeek-V3.1 | 60 | 141,204 | -| DeepSeek-R1 | 60 | 129,883 | -| Gemini-2.5 | 60 | 91,585 | -| Gemini-2.5-NT | 60 | 65,826 | -| Qwen3-235b | 60 | 76,512 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 117,358 | -| GPT-4o-mini | 60 | 102,564 | -| DeepSeek-V3.1 | 60 | 141,204 | -| DeepSeek-R1 | 59 | 128,948 | -| Gemini-2.5 | 60 | 91,585 | -| Gemini-2.5-NT | 60 | 65,826 | -| Qwen3-235b | 60 | 76,512 | - -### Pass (with retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 1 | 185,051 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 0 | - | - -### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): Gemini-2.5-NT = 65,826 tokens -- Maximum (highest mean): DeepSeek-V3.1 = 141,204 tokens -- Delta: 75,378 (114.5%) - -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): Gemini-2.5-NT = 65,826 tokens -- Maximum (highest mean): DeepSeek-V3.1 = 141,204 tokens -- Delta: 75,378 (114.5%) - -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): DeepSeek-R1 = 185,051 tokens -- Maximum (highest mean): DeepSeek-R1 = 185,051 tokens -- Delta: 0 (0.0%) - -### Baseline vs maximum — Failure -- No data to compare. - -## write_a_book_with_flows-MCP - -### Overall per-model total tokens (all statuses) -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 190,842 | -| GPT-4o-mini | 60 | 133,339 | -| DeepSeek-V3.1 | 60 | 184,315 | -| DeepSeek-R1 | 60 | 190,705 | -| Gemini-2.5 | 60 | 107,154 | -| Gemini-2.5-NT | 60 | 82,546 | -| Qwen3-235b | 60 | 110,716 | - -### Pass (no retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 190,842 | -| GPT-4o-mini | 60 | 133,339 | -| DeepSeek-V3.1 | 60 | 184,315 | -| DeepSeek-R1 | 52 | 171,033 | -| Gemini-2.5 | 60 | 107,154 | -| Gemini-2.5-NT | 59 | 82,348 | -| Qwen3-235b | 59 | 110,871 | - -### Pass (with retries) distribution -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 8 | 318,572 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 1 | 94,226 | -| Qwen3-235b | 1 | 101,570 | - -### Baseline vs maximum (overall mean tokens) -- Baseline (lowest mean): Gemini-2.5-NT = 82,546 tokens -- Maximum (highest mean): GPT-5 = 190,842 tokens -- Delta: 108,296 (131.2%) - -### Baseline vs maximum — Pass (no retries) -- Baseline (lowest mean): Gemini-2.5-NT = 82,348 tokens -- Maximum (highest mean): GPT-5 = 190,842 tokens -- Delta: 108,494 (131.8%) - -### Baseline vs maximum — Pass (with retries) -- Baseline (lowest mean): Gemini-2.5-NT = 94,226 tokens -- Maximum (highest mean): DeepSeek-R1 = 318,572 tokens -- Delta: 224,346 (238.1%) - -### Baseline vs maximum — Failure -- No data to compare. \ No newline at end of file +- Delta: 295,505 (714.5%) \ No newline at end of file diff --git a/data/processed/RQ3/Violin/project_token_stats_pass_vs_failure.md b/data/processed/RQ3/Violin/project_token_stats_pass_vs_failure.md index 46597daf9969661cb5bdad99e5e293fd6d39a963..6f94aa7c6c5001d525cd761aa37f9207003d1a7f 100644 --- a/data/processed/RQ3/Violin/project_token_stats_pass_vs_failure.md +++ b/data/processed/RQ3/Violin/project_token_stats_pass_vs_failure.md @@ -4,152 +4,90 @@ | Status | n | Mean | | --- | --- | --- | -| Pass (no retries) | 6849 | 47,691 | +| Direct Success | 6849 | 47,691 | | Failure | 862 | 29,935 | -## SQL_assistant-A2A +## BookWriter-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 36 | 8,736 | -| GPT-4o-mini | 50 | 17,092 | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 55,457 | -| Gemini-2.5 | 2 | 23,118 | -| Gemini-2.5-NT | 2 | 36,636 | -| Qwen3-235b | 35 | 13,848 | - -### Failure - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 38 | 26,988 | -| Gemini-2.5-NT | 53 | 21,051 | -| Qwen3-235b | 2 | 9,587 | - -### Pass (no retries) vs Failure (mean, abs & %) - -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 8,736 | - | - | -| GPT-4o-mini | 17,092 | - | - | -| DeepSeek-V3.1 | - | - | - | -| DeepSeek-R1 | 55,457 | - | - | -| Gemini-2.5 | 23,118 | 26,988 | +3,870 (16.7%) | -| Gemini-2.5-NT | 36,636 | 21,051 | -15,585 (-42.5%) | -| Qwen3-235b | 13,848 | 9,587 | -4,261 (-30.8%) | - -## SQL_assistant-A2A_mix - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 46 | 6,953 | -| GPT-4o-mini | 43 | 10,004 | -| DeepSeek-V3.1 | 44 | 12,571 | -| DeepSeek-R1 | 20 | 29,145 | -| Gemini-2.5 | 25 | 14,332 | -| Gemini-2.5-NT | 33 | 9,561 | -| Qwen3-235b | 40 | 10,594 | - -### Failure - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 9 | 35,120 | -| Gemini-2.5-NT | 8 | 19,462 | -| Qwen3-235b | 0 | - | +| GPT-5 | 60 | 190,842 | +| GPT-4o-mini | 60 | 133,339 | +| DeepSeek-V3.1 | 60 | 184,315 | +| DeepSeek-R1 | 52 | 171,033 | +| Gemini-2.5 | 60 | 107,154 | +| Gemini-2.5-NT | 59 | 82,348 | +| Qwen3-235b | 59 | 110,871 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 6,953 | - | - | -| GPT-4o-mini | 10,004 | - | - | -| DeepSeek-V3.1 | 12,571 | - | - | -| DeepSeek-R1 | 29,145 | - | - | -| Gemini-2.5 | 14,332 | 35,120 | +20,788 (145.0%) | -| Gemini-2.5-NT | 9,561 | 19,462 | +9,902 (103.6%) | -| Qwen3-235b | 10,594 | - | - | - -## SQL_assistant-MCP - -### Pass (no retries) +| GPT-5 | 190,842 | - | - | +| GPT-4o-mini | 133,339 | - | - | +| DeepSeek-V3.1 | 184,315 | - | - | +| DeepSeek-R1 | 171,033 | - | - | +| Gemini-2.5 | 107,154 | - | - | +| Gemini-2.5-NT | 82,348 | - | - | +| Qwen3-235b | 110,871 | - | - | -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 19 | 6,378 | -| GPT-4o-mini | 21 | 14,334 | -| DeepSeek-V3.1 | 26 | 15,771 | -| DeepSeek-R1 | 13 | 35,319 | -| Gemini-2.5 | 6 | 15,321 | -| Gemini-2.5-NT | 10 | 12,100 | -| Qwen3-235b | 37 | 12,304 | +## BookWriter-A2A -### Failure +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 43 | 13,573 | -| Gemini-2.5-NT | 44 | 13,166 | -| Qwen3-235b | 0 | - | +| GPT-5 | 60 | 181,022 | +| GPT-4o-mini | 60 | 139,665 | +| DeepSeek-V3.1 | 59 | 270,142 | +| DeepSeek-R1 | 52 | 173,528 | +| Gemini-2.5 | 60 | 110,126 | +| Gemini-2.5-NT | 60 | 86,249 | +| Qwen3-235b | 60 | 118,406 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 6,378 | - | - | -| GPT-4o-mini | 14,334 | - | - | -| DeepSeek-V3.1 | 15,771 | - | - | -| DeepSeek-R1 | 35,319 | - | - | -| Gemini-2.5 | 15,321 | 13,573 | -1,748 (-11.4%) | -| Gemini-2.5-NT | 12,100 | 13,166 | +1,066 (8.8%) | -| Qwen3-235b | 12,304 | - | - | +| GPT-5 | 181,022 | - | - | +| GPT-4o-mini | 139,665 | - | - | +| DeepSeek-V3.1 | 270,142 | - | - | +| DeepSeek-R1 | 173,528 | - | - | +| Gemini-2.5 | 110,126 | - | - | +| Gemini-2.5-NT | 86,249 | - | - | +| Qwen3-235b | 118,406 | - | - | -## email_auto_responder_flow-MCP +## BookWriter-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 100 | 6,903 | -| GPT-4o-mini | 100 | 9,437 | -| DeepSeek-V3.1 | 98 | 9,531 | -| DeepSeek-R1 | 96 | 14,798 | -| Gemini-2.5 | 91 | 33,629 | -| Gemini-2.5-NT | 88 | 12,696 | -| Qwen3-235b | 100 | 7,434 | +| GPT-5 | 60 | 117,358 | +| GPT-4o-mini | 60 | 102,564 | +| DeepSeek-V3.1 | 60 | 141,204 | +| DeepSeek-R1 | 59 | 128,948 | +| Gemini-2.5 | 60 | 91,585 | +| Gemini-2.5-NT | 60 | 65,826 | +| Qwen3-235b | 60 | 76,512 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 6,903 | - | - | -| GPT-4o-mini | 9,437 | - | - | -| DeepSeek-V3.1 | 9,531 | - | - | -| DeepSeek-R1 | 14,798 | - | - | -| Gemini-2.5 | 33,629 | - | - | -| Gemini-2.5-NT | 12,696 | - | - | -| Qwen3-235b | 7,434 | - | - | +| GPT-5 | 117,358 | - | - | +| GPT-4o-mini | 102,564 | - | - | +| DeepSeek-V3.1 | 141,204 | - | - | +| DeepSeek-R1 | 128,948 | - | - | +| Gemini-2.5 | 91,585 | - | - | +| Gemini-2.5-NT | 65,826 | - | - | +| Qwen3-235b | 76,512 | - | - | -## email_auto_responder_flow +## EmailResponder -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -173,9 +111,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 0 | - | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 5,562 | - | - | | GPT-4o-mini | 8,044 | - | - | @@ -185,35 +123,35 @@ | Gemini-2.5-NT | 12,494 | - | - | | Qwen3-235b | 5,900 | - | - | -## game_builder-MCP +## EmailResponder-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 25 | 22,425 | -| GPT-4o-mini | 25 | 47,996 | -| DeepSeek-V3.1 | 25 | 55,289 | -| DeepSeek-R1 | 25 | 44,851 | -| Gemini-2.5 | 25 | 75,519 | -| Gemini-2.5-NT | 25 | 50,782 | -| Qwen3-235b | 25 | 24,039 | +| GPT-5 | 100 | 6,903 | +| GPT-4o-mini | 100 | 9,437 | +| DeepSeek-V3.1 | 98 | 9,531 | +| DeepSeek-R1 | 96 | 14,798 | +| Gemini-2.5 | 91 | 33,629 | +| Gemini-2.5-NT | 88 | 12,696 | +| Qwen3-235b | 100 | 7,434 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 22,425 | - | - | -| GPT-4o-mini | 47,996 | - | - | -| DeepSeek-V3.1 | 55,289 | - | - | -| DeepSeek-R1 | 44,851 | - | - | -| Gemini-2.5 | 75,519 | - | - | -| Gemini-2.5-NT | 50,782 | - | - | -| Qwen3-235b | 24,039 | - | - | +| GPT-5 | 6,903 | - | - | +| GPT-4o-mini | 9,437 | - | - | +| DeepSeek-V3.1 | 9,531 | - | - | +| DeepSeek-R1 | 14,798 | - | - | +| Gemini-2.5 | 33,629 | - | - | +| Gemini-2.5-NT | 12,696 | - | - | +| Qwen3-235b | 7,434 | - | - | -## game_builder +## GameBuilder -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -237,9 +175,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 0 | - | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 22,904 | - | - | | GPT-4o-mini | 17,823 | - | - | @@ -249,111 +187,73 @@ | Gemini-2.5-NT | 54,423 | - | - | | Qwen3-235b | 24,212 | - | - | -## intelligent_recruitment_platform-A2A - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 31,991 | -| GPT-4o-mini | 60 | 74,951 | -| DeepSeek-V3.1 | 60 | 65,662 | -| DeepSeek-R1 | 56 | 73,964 | -| Gemini-2.5 | 60 | 68,719 | -| Gemini-2.5-NT | 60 | 55,660 | -| Qwen3-235b | 44 | 73,316 | +## GameBuilder-MCP -### Failure +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 1 | 32,043 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 1 | 169,975 | - -### Pass (no retries) vs Failure (mean, abs & %) - -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 31,991 | - | - | -| GPT-4o-mini | 74,951 | - | - | -| DeepSeek-V3.1 | 65,662 | - | - | -| DeepSeek-R1 | 73,964 | 32,043 | -41,921 (-56.7%) | -| Gemini-2.5 | 68,719 | - | - | -| Gemini-2.5-NT | 55,660 | - | - | -| Qwen3-235b | 73,316 | 169,975 | +96,659 (131.8%) | - -## intelligent_recruitment_platform-A2A_mix - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 32,672 | -| GPT-4o-mini | 60 | 101,936 | -| DeepSeek-V3.1 | 60 | 65,048 | -| DeepSeek-R1 | 57 | 76,741 | -| Gemini-2.5 | 60 | 57,128 | -| Gemini-2.5-NT | 60 | 36,887 | -| Qwen3-235b | 49 | 67,089 | +| GPT-5 | 25 | 22,425 | +| GPT-4o-mini | 25 | 47,996 | +| DeepSeek-V3.1 | 25 | 55,289 | +| DeepSeek-R1 | 25 | 44,851 | +| Gemini-2.5 | 25 | 75,519 | +| Gemini-2.5-NT | 25 | 50,782 | +| Qwen3-235b | 25 | 24,039 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 32,672 | - | - | -| GPT-4o-mini | 101,936 | - | - | -| DeepSeek-V3.1 | 65,048 | - | - | -| DeepSeek-R1 | 76,741 | - | - | -| Gemini-2.5 | 57,128 | - | - | -| Gemini-2.5-NT | 36,887 | - | - | -| Qwen3-235b | 67,089 | - | - | +| GPT-5 | 22,425 | - | - | +| GPT-4o-mini | 47,996 | - | - | +| DeepSeek-V3.1 | 55,289 | - | - | +| DeepSeek-R1 | 44,851 | - | - | +| Gemini-2.5 | 75,519 | - | - | +| Gemini-2.5-NT | 50,782 | - | - | +| Qwen3-235b | 24,039 | - | - | -## intelligent_recruitment_platform-MCP +## LandingPageGenerator-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 32,411 | -| GPT-4o-mini | 60 | 99,933 | -| DeepSeek-V3.1 | 60 | 65,748 | -| DeepSeek-R1 | 58 | 82,498 | -| Gemini-2.5 | 60 | 64,864 | -| Gemini-2.5-NT | 60 | 54,402 | -| Qwen3-235b | 40 | 68,264 | +| GPT-5 | 60 | 12,653 | +| GPT-4o-mini | 59 | 19,333 | +| DeepSeek-V3.1 | 55 | 15,912 | +| DeepSeek-R1 | 33 | 24,081 | +| Gemini-2.5 | 1 | 55,158 | +| Gemini-2.5-NT | 2 | 188,992 | +| Qwen3-235b | 53 | 15,068 | ### Failure | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 2 | 128,765 | +| GPT-4o-mini | 1 | 14,308 | +| DeepSeek-V3.1 | 5 | 15,362 | +| DeepSeek-R1 | 19 | 21,990 | +| Gemini-2.5 | 59 | 26,041 | +| Gemini-2.5-NT | 58 | 28,064 | +| Qwen3-235b | 7 | 10,009 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 32,411 | - | - | -| GPT-4o-mini | 99,933 | - | - | -| DeepSeek-V3.1 | 65,748 | - | - | -| DeepSeek-R1 | 82,498 | - | - | -| Gemini-2.5 | 64,864 | - | - | -| Gemini-2.5-NT | 54,402 | - | - | -| Qwen3-235b | 68,264 | 128,765 | +60,501 (88.6%) | +| GPT-5 | 12,653 | - | - | +| GPT-4o-mini | 19,333 | 14,308 | -5,025 (-26.0%) | +| DeepSeek-V3.1 | 15,912 | 15,362 | -550 (-3.5%) | +| DeepSeek-R1 | 24,081 | 21,990 | -2,091 (-8.7%) | +| Gemini-2.5 | 55,158 | 26,041 | -29,117 (-52.8%) | +| Gemini-2.5-NT | 188,992 | 28,064 | -160,929 (-85.2%) | +| Qwen3-235b | 15,068 | 10,009 | -5,059 (-33.6%) | -## landing_page_generator-A2A +## LandingPageGenerator-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -377,9 +277,9 @@ | Gemini-2.5-NT | 60 | 22,826 | | Qwen3-235b | 30 | 14,804 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 18,863 | 13,548 | -5,315 (-28.2%) | | GPT-4o-mini | 20,076 | 27,976 | +7,900 (39.4%) | @@ -389,9 +289,9 @@ | Gemini-2.5-NT | - | 22,826 | - | | Qwen3-235b | 20,066 | 14,804 | -5,263 (-26.2%) | -## landing_page_generator-A2A_mix +## LandingPageGenerator-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -415,9 +315,9 @@ | Gemini-2.5-NT | 60 | 14,949 | | Qwen3-235b | 29 | 14,748 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 29,492 | 22,299 | -7,194 (-24.4%) | | GPT-4o-mini | 43,589 | - | - | @@ -427,47 +327,35 @@ | Gemini-2.5-NT | - | 14,949 | - | | Qwen3-235b | 19,428 | 14,748 | -4,680 (-24.1%) | -## landing_page_generator-MCP - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 12,653 | -| GPT-4o-mini | 59 | 19,333 | -| DeepSeek-V3.1 | 55 | 15,912 | -| DeepSeek-R1 | 33 | 24,081 | -| Gemini-2.5 | 1 | 55,158 | -| Gemini-2.5-NT | 2 | 188,992 | -| Qwen3-235b | 53 | 15,068 | +## MarkdownValidator -### Failure +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 1 | 14,308 | -| DeepSeek-V3.1 | 5 | 15,362 | -| DeepSeek-R1 | 19 | 21,990 | -| Gemini-2.5 | 59 | 26,041 | -| Gemini-2.5-NT | 58 | 28,064 | -| Qwen3-235b | 7 | 10,009 | +| GPT-5 | 60 | 1,853 | +| GPT-4o-mini | 60 | 1,853 | +| DeepSeek-V3.1 | 60 | 1,948 | +| DeepSeek-R1 | 60 | 5,912 | +| Gemini-2.5 | 47 | 81,437 | +| Gemini-2.5-NT | 57 | 16,246 | +| Qwen3-235b | 60 | 1,854 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 12,653 | - | - | -| GPT-4o-mini | 19,333 | 14,308 | -5,025 (-26.0%) | -| DeepSeek-V3.1 | 15,912 | 15,362 | -550 (-3.5%) | -| DeepSeek-R1 | 24,081 | 21,990 | -2,091 (-8.7%) | -| Gemini-2.5 | 55,158 | 26,041 | -29,117 (-52.8%) | -| Gemini-2.5-NT | 188,992 | 28,064 | -160,929 (-85.2%) | -| Qwen3-235b | 15,068 | 10,009 | -5,059 (-33.6%) | +| GPT-5 | 1,853 | - | - | +| GPT-4o-mini | 1,853 | - | - | +| DeepSeek-V3.1 | 1,948 | - | - | +| DeepSeek-R1 | 5,912 | - | - | +| Gemini-2.5 | 81,437 | - | - | +| Gemini-2.5-NT | 16,246 | - | - | +| Qwen3-235b | 1,854 | - | - | -## markdown_validator-MCP +## MarkdownValidator-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -479,9 +367,9 @@ | Gemini-2.5-NT | 56 | 39,448 | | Qwen3-235b | 60 | 1,994 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 1,967 | - | - | | GPT-4o-mini | 1,942 | - | - | @@ -491,83 +379,121 @@ | Gemini-2.5-NT | 39,448 | - | - | | Qwen3-235b | 1,994 | - | - | -## markdown_validator +## RecruitmentAssistant-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 1,853 | -| GPT-4o-mini | 60 | 1,853 | -| DeepSeek-V3.1 | 60 | 1,948 | -| DeepSeek-R1 | 60 | 5,912 | -| Gemini-2.5 | 47 | 81,437 | -| Gemini-2.5-NT | 57 | 16,246 | -| Qwen3-235b | 60 | 1,854 | +| GPT-5 | 60 | 32,411 | +| GPT-4o-mini | 60 | 99,933 | +| DeepSeek-V3.1 | 60 | 65,748 | +| DeepSeek-R1 | 58 | 82,498 | +| Gemini-2.5 | 60 | 64,864 | +| Gemini-2.5-NT | 60 | 54,402 | +| Qwen3-235b | 40 | 68,264 | + +### Failure + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 2 | 128,765 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 1,853 | - | - | -| GPT-4o-mini | 1,853 | - | - | -| DeepSeek-V3.1 | 1,948 | - | - | -| DeepSeek-R1 | 5,912 | - | - | -| Gemini-2.5 | 81,437 | - | - | -| Gemini-2.5-NT | 16,246 | - | - | -| Qwen3-235b | 1,854 | - | - | +| GPT-5 | 32,411 | - | - | +| GPT-4o-mini | 99,933 | - | - | +| DeepSeek-V3.1 | 65,748 | - | - | +| DeepSeek-R1 | 82,498 | - | - | +| Gemini-2.5 | 64,864 | - | - | +| Gemini-2.5-NT | 54,402 | - | - | +| Qwen3-235b | 68,264 | 128,765 | +60,501 (88.6%) | -## self_evaluation_loop_flow-A2A +## RecruitmentAssistant-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 31 | 23,996 | -| GPT-4o-mini | 24 | 42,258 | -| DeepSeek-V3.1 | 46 | 70,407 | -| DeepSeek-R1 | 21 | 45,776 | -| Gemini-2.5 | 56 | 12,176 | -| Gemini-2.5-NT | 53 | 10,675 | -| Qwen3-235b | 27 | 30,147 | +| GPT-5 | 60 | 31,991 | +| GPT-4o-mini | 60 | 74,951 | +| DeepSeek-V3.1 | 60 | 65,662 | +| DeepSeek-R1 | 56 | 73,964 | +| Gemini-2.5 | 60 | 68,719 | +| Gemini-2.5-NT | 60 | 55,660 | +| Qwen3-235b | 44 | 73,316 | ### Failure | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 2 | 98,808 | -| GPT-4o-mini | 4 | 213,454 | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 5 | 181,907 | +| DeepSeek-R1 | 1 | 32,043 | | Gemini-2.5 | 0 | - | | Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 1 | 230,991 | +| Qwen3-235b | 1 | 169,975 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 23,996 | 98,808 | +74,812 (311.8%) | -| GPT-4o-mini | 42,258 | 213,454 | +171,196 (405.1%) | -| DeepSeek-V3.1 | 70,407 | - | - | -| DeepSeek-R1 | 45,776 | 181,907 | +136,131 (297.4%) | -| Gemini-2.5 | 12,176 | - | - | -| Gemini-2.5-NT | 10,675 | - | - | -| Qwen3-235b | 30,147 | 230,991 | +200,844 (666.2%) | +| GPT-5 | 31,991 | - | - | +| GPT-4o-mini | 74,951 | - | - | +| DeepSeek-V3.1 | 65,662 | - | - | +| DeepSeek-R1 | 73,964 | 32,043 | -41,921 (-56.7%) | +| Gemini-2.5 | 68,719 | - | - | +| Gemini-2.5-NT | 55,660 | - | - | +| Qwen3-235b | 73,316 | 169,975 | +96,659 (131.8%) | -## self_evaluation_loop_flow-A2A_mix +## RecruitmentAssistant-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 55 | 15,857 | -| GPT-4o-mini | 51 | 33,008 | -| DeepSeek-V3.1 | 49 | 51,074 | -| DeepSeek-R1 | 27 | 40,199 | -| Gemini-2.5 | 35 | 13,495 | -| Gemini-2.5-NT | 45 | 7,487 | -| Qwen3-235b | 35 | 14,019 | +| GPT-5 | 60 | 32,672 | +| GPT-4o-mini | 60 | 101,936 | +| DeepSeek-V3.1 | 60 | 65,048 | +| DeepSeek-R1 | 57 | 76,741 | +| Gemini-2.5 | 60 | 57,128 | +| Gemini-2.5-NT | 60 | 36,887 | +| Qwen3-235b | 49 | 67,089 | + +### Direct Success vs Failure (mean, abs & %) + +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 32,672 | - | - | +| GPT-4o-mini | 101,936 | - | - | +| DeepSeek-V3.1 | 65,048 | - | - | +| DeepSeek-R1 | 76,741 | - | - | +| Gemini-2.5 | 57,128 | - | - | +| Gemini-2.5-NT | 36,887 | - | - | +| Qwen3-235b | 67,089 | - | - | + +## SQLAssistant-MCP + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 19 | 6,378 | +| GPT-4o-mini | 21 | 14,334 | +| DeepSeek-V3.1 | 26 | 15,771 | +| DeepSeek-R1 | 13 | 35,319 | +| Gemini-2.5 | 6 | 15,321 | +| Gemini-2.5-NT | 10 | 12,100 | +| Qwen3-235b | 37 | 12,304 | ### Failure @@ -576,26 +502,102 @@ | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 2 | 336,866 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 1 | 41,361 | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 43 | 13,573 | +| Gemini-2.5-NT | 44 | 13,166 | +| Qwen3-235b | 0 | - | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 15,857 | - | - | -| GPT-4o-mini | 33,008 | - | - | -| DeepSeek-V3.1 | 51,074 | - | - | -| DeepSeek-R1 | 40,199 | 336,866 | +296,667 (738.0%) | -| Gemini-2.5 | 13,495 | - | - | -| Gemini-2.5-NT | 7,487 | - | - | -| Qwen3-235b | 14,019 | 41,361 | +27,342 (195.0%) | +| GPT-5 | 6,378 | - | - | +| GPT-4o-mini | 14,334 | - | - | +| DeepSeek-V3.1 | 15,771 | - | - | +| DeepSeek-R1 | 35,319 | - | - | +| Gemini-2.5 | 15,321 | 13,573 | -1,748 (-11.4%) | +| Gemini-2.5-NT | 12,100 | 13,166 | +1,066 (8.8%) | +| Qwen3-235b | 12,304 | - | - | + +## SQLAssistant-A2A + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 36 | 8,736 | +| GPT-4o-mini | 50 | 17,092 | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 3 | 55,457 | +| Gemini-2.5 | 2 | 23,118 | +| Gemini-2.5-NT | 2 | 36,636 | +| Qwen3-235b | 35 | 13,848 | + +### Failure + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 38 | 26,988 | +| Gemini-2.5-NT | 53 | 21,051 | +| Qwen3-235b | 2 | 9,587 | + +### Direct Success vs Failure (mean, abs & %) + +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 8,736 | - | - | +| GPT-4o-mini | 17,092 | - | - | +| DeepSeek-V3.1 | - | - | - | +| DeepSeek-R1 | 55,457 | - | - | +| Gemini-2.5 | 23,118 | 26,988 | +3,870 (16.7%) | +| Gemini-2.5-NT | 36,636 | 21,051 | -15,585 (-42.5%) | +| Qwen3-235b | 13,848 | 9,587 | -4,261 (-30.8%) | + +## SQLAssistant-H-A2A + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 46 | 6,953 | +| GPT-4o-mini | 43 | 10,004 | +| DeepSeek-V3.1 | 44 | 12,571 | +| DeepSeek-R1 | 20 | 29,145 | +| Gemini-2.5 | 25 | 14,332 | +| Gemini-2.5-NT | 33 | 9,561 | +| Qwen3-235b | 40 | 10,594 | + +### Failure + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 9 | 35,120 | +| Gemini-2.5-NT | 8 | 19,462 | +| Qwen3-235b | 0 | - | + +### Direct Success vs Failure (mean, abs & %) + +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 6,953 | - | - | +| GPT-4o-mini | 10,004 | - | - | +| DeepSeek-V3.1 | 12,571 | - | - | +| DeepSeek-R1 | 29,145 | - | - | +| Gemini-2.5 | 14,332 | 35,120 | +20,788 (145.0%) | +| Gemini-2.5-NT | 9,561 | 19,462 | +9,902 (103.6%) | +| Qwen3-235b | 10,594 | - | - | -## self_evaluation_loop_flow-MCP +## SocialMediaManager-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -619,9 +621,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 1 | 110,239 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 22,287 | 99,433 | +77,146 (346.1%) | | GPT-4o-mini | 40,555 | 227,117 | +186,562 (460.0%) | @@ -631,127 +633,113 @@ | Gemini-2.5-NT | 10,904 | - | - | | Qwen3-235b | 27,216 | 110,239 | +83,023 (305.1%) | -## write_a_book_with_flows-A2A +## SocialMediaManager-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 181,022 | -| GPT-4o-mini | 60 | 139,665 | -| DeepSeek-V3.1 | 59 | 270,142 | -| DeepSeek-R1 | 52 | 173,528 | -| Gemini-2.5 | 60 | 110,126 | -| Gemini-2.5-NT | 60 | 86,249 | -| Qwen3-235b | 60 | 118,406 | - -### Pass (no retries) vs Failure (mean, abs & %) - -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 181,022 | - | - | -| GPT-4o-mini | 139,665 | - | - | -| DeepSeek-V3.1 | 270,142 | - | - | -| DeepSeek-R1 | 173,528 | - | - | -| Gemini-2.5 | 110,126 | - | - | -| Gemini-2.5-NT | 86,249 | - | - | -| Qwen3-235b | 118,406 | - | - | - -## write_a_book_with_flows-A2A_mix +| GPT-5 | 31 | 23,996 | +| GPT-4o-mini | 24 | 42,258 | +| DeepSeek-V3.1 | 46 | 70,407 | +| DeepSeek-R1 | 21 | 45,776 | +| Gemini-2.5 | 56 | 12,176 | +| Gemini-2.5-NT | 53 | 10,675 | +| Qwen3-235b | 27 | 30,147 | -### Pass (no retries) +### Failure | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 117,358 | -| GPT-4o-mini | 60 | 102,564 | -| DeepSeek-V3.1 | 60 | 141,204 | -| DeepSeek-R1 | 59 | 128,948 | -| Gemini-2.5 | 60 | 91,585 | -| Gemini-2.5-NT | 60 | 65,826 | -| Qwen3-235b | 60 | 76,512 | +| GPT-5 | 2 | 98,808 | +| GPT-4o-mini | 4 | 213,454 | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 5 | 181,907 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 1 | 230,991 | -### Pass (no retries) vs Failure (mean, abs & %) +### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 117,358 | - | - | -| GPT-4o-mini | 102,564 | - | - | -| DeepSeek-V3.1 | 141,204 | - | - | -| DeepSeek-R1 | 128,948 | - | - | -| Gemini-2.5 | 91,585 | - | - | -| Gemini-2.5-NT | 65,826 | - | - | -| Qwen3-235b | 76,512 | - | - | +| GPT-5 | 23,996 | 98,808 | +74,812 (311.8%) | +| GPT-4o-mini | 42,258 | 213,454 | +171,196 (405.1%) | +| DeepSeek-V3.1 | 70,407 | - | - | +| DeepSeek-R1 | 45,776 | 181,907 | +136,131 (297.4%) | +| Gemini-2.5 | 12,176 | - | - | +| Gemini-2.5-NT | 10,675 | - | - | +| Qwen3-235b | 30,147 | 230,991 | +200,844 (666.2%) | -## write_a_book_with_flows-MCP +## SocialMediaManager-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 190,842 | -| GPT-4o-mini | 60 | 133,339 | -| DeepSeek-V3.1 | 60 | 184,315 | -| DeepSeek-R1 | 52 | 171,033 | -| Gemini-2.5 | 60 | 107,154 | -| Gemini-2.5-NT | 59 | 82,348 | -| Qwen3-235b | 59 | 110,871 | +| GPT-5 | 55 | 15,857 | +| GPT-4o-mini | 51 | 33,008 | +| DeepSeek-V3.1 | 49 | 51,074 | +| DeepSeek-R1 | 27 | 40,199 | +| Gemini-2.5 | 35 | 13,495 | +| Gemini-2.5-NT | 45 | 7,487 | +| Qwen3-235b | 35 | 14,019 | -### Pass (no retries) vs Failure (mean, abs & %) +### Failure -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 190,842 | - | - | -| GPT-4o-mini | 133,339 | - | - | -| DeepSeek-V3.1 | 184,315 | - | - | -| DeepSeek-R1 | 171,033 | - | - | -| Gemini-2.5 | 107,154 | - | - | -| Gemini-2.5-NT | 82,348 | - | - | -| Qwen3-235b | 110,871 | - | - | +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 2 | 336,866 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 1 | 41,361 | -## Series aggregates (by task prefix) +### Direct Success vs Failure (mean, abs & %) -### SQL_assistant (aggregated across architectures) +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 15,857 | - | - | +| GPT-4o-mini | 33,008 | - | - | +| DeepSeek-V3.1 | 51,074 | - | - | +| DeepSeek-R1 | 40,199 | 336,866 | +296,667 (738.0%) | +| Gemini-2.5 | 13,495 | - | - | +| Gemini-2.5-NT | 7,487 | - | - | +| Qwen3-235b | 14,019 | 41,361 | +27,342 (195.0%) | -### Pass (no retries) +## Series aggregates (by task prefix) -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 101 | 7,480 | -| GPT-4o-mini | 114 | 13,910 | -| DeepSeek-V3.1 | 70 | 13,760 | -| DeepSeek-R1 | 36 | 33,567 | -| Gemini-2.5 | 33 | 15,045 | -| Gemini-2.5-NT | 45 | 11,328 | -| Qwen3-235b | 112 | 12,176 | +### BookWriter (aggregated across variants) -### Failure +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 90 | 21,392 | -| Gemini-2.5-NT | 105 | 17,626 | -| Qwen3-235b | 2 | 9,587 | +| GPT-5 | 180 | 163,074 | +| GPT-4o-mini | 180 | 125,190 | +| DeepSeek-V3.1 | 179 | 198,154 | +| DeepSeek-R1 | 163 | 156,596 | +| Gemini-2.5 | 180 | 102,955 | +| Gemini-2.5-NT | 179 | 78,117 | +| Qwen3-235b | 179 | 101,880 | -#### Pass (no retries) vs Failure (mean, abs & %) +#### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 7,480 | - | - | -| GPT-4o-mini | 13,910 | - | - | -| DeepSeek-V3.1 | 13,760 | - | - | -| DeepSeek-R1 | 33,567 | - | - | -| Gemini-2.5 | 15,045 | 21,392 | +6,347 (42.2%) | -| Gemini-2.5-NT | 11,328 | 17,626 | +6,297 (55.6%) | -| Qwen3-235b | 12,176 | 9,587 | -2,589 (-21.3%) | +| GPT-5 | 163,074 | - | - | +| GPT-4o-mini | 125,190 | - | - | +| DeepSeek-V3.1 | 198,154 | - | - | +| DeepSeek-R1 | 156,596 | - | - | +| Gemini-2.5 | 102,955 | - | - | +| Gemini-2.5-NT | 78,117 | - | - | +| Qwen3-235b | 101,880 | - | - | -### email_auto_responder_flow (aggregated across architectures) +### EmailResponder (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -775,9 +763,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 0 | - | -#### Pass (no retries) vs Failure (mean, abs & %) +#### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 6,232 | - | - | | GPT-4o-mini | 8,741 | - | - | @@ -787,9 +775,9 @@ | Gemini-2.5-NT | 12,595 | - | - | | Qwen3-235b | 6,667 | - | - | -### game_builder (aggregated across architectures) +### GameBuilder (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -813,9 +801,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 0 | - | -#### Pass (no retries) vs Failure (mean, abs & %) +#### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 22,665 | - | - | | GPT-4o-mini | 32,909 | - | - | @@ -825,47 +813,9 @@ | Gemini-2.5-NT | 52,527 | - | - | | Qwen3-235b | 24,125 | - | - | -### intelligent_recruitment_platform (aggregated across architectures) - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 180 | 32,358 | -| GPT-4o-mini | 180 | 92,274 | -| DeepSeek-V3.1 | 180 | 65,486 | -| DeepSeek-R1 | 171 | 77,784 | -| Gemini-2.5 | 180 | 63,570 | -| Gemini-2.5-NT | 180 | 48,983 | -| Qwen3-235b | 133 | 69,503 | - -### Failure - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 1 | 32,043 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 3 | 142,502 | - -#### Pass (no retries) vs Failure (mean, abs & %) - -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 32,358 | - | - | -| GPT-4o-mini | 92,274 | - | - | -| DeepSeek-V3.1 | 65,486 | - | - | -| DeepSeek-R1 | 77,784 | 32,043 | -45,741 (-58.8%) | -| Gemini-2.5 | 63,570 | - | - | -| Gemini-2.5-NT | 48,983 | - | - | -| Qwen3-235b | 69,503 | 142,502 | +72,999 (105.0%) | - -### landing_page_generator (aggregated across architectures) +### LandingPageGenerator (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -889,9 +839,9 @@ | Gemini-2.5-NT | 178 | 21,877 | | Qwen3-235b | 66 | 14,271 | -#### Pass (no retries) vs Failure (mean, abs & %) +#### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 17,805 | 17,425 | -381 (-2.1%) | | GPT-4o-mini | 27,799 | 23,420 | -4,379 (-15.8%) | @@ -901,9 +851,9 @@ | Gemini-2.5-NT | 188,992 | 21,877 | -167,115 (-88.4%) | | Qwen3-235b | 17,466 | 14,271 | -3,195 (-18.3%) | -### markdown_validator (aggregated across architectures) +### MarkdownValidator (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -915,9 +865,9 @@ | Gemini-2.5-NT | 113 | 27,744 | | Qwen3-235b | 120 | 1,924 | -#### Pass (no retries) vs Failure (mean, abs & %) +#### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 1,910 | - | - | | GPT-4o-mini | 1,897 | - | - | @@ -927,9 +877,85 @@ | Gemini-2.5-NT | 27,744 | - | - | | Qwen3-235b | 1,924 | - | - | -### self_evaluation_loop_flow (aggregated across architectures) +### RecruitmentAssistant (aggregated across variants) + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 180 | 32,358 | +| GPT-4o-mini | 180 | 92,274 | +| DeepSeek-V3.1 | 180 | 65,486 | +| DeepSeek-R1 | 171 | 77,784 | +| Gemini-2.5 | 180 | 63,570 | +| Gemini-2.5-NT | 180 | 48,983 | +| Qwen3-235b | 133 | 69,503 | + +### Failure + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 1 | 32,043 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 3 | 142,502 | + +#### Direct Success vs Failure (mean, abs & %) + +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 32,358 | - | - | +| GPT-4o-mini | 92,274 | - | - | +| DeepSeek-V3.1 | 65,486 | - | - | +| DeepSeek-R1 | 77,784 | 32,043 | -45,741 (-58.8%) | +| Gemini-2.5 | 63,570 | - | - | +| Gemini-2.5-NT | 48,983 | - | - | +| Qwen3-235b | 69,503 | 142,502 | +72,999 (105.0%) | + +### SQLAssistant (aggregated across variants) + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 101 | 7,480 | +| GPT-4o-mini | 114 | 13,910 | +| DeepSeek-V3.1 | 70 | 13,760 | +| DeepSeek-R1 | 36 | 33,567 | +| Gemini-2.5 | 33 | 15,045 | +| Gemini-2.5-NT | 45 | 11,328 | +| Qwen3-235b | 112 | 12,176 | + +### Failure + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 90 | 21,392 | +| Gemini-2.5-NT | 105 | 17,626 | +| Qwen3-235b | 2 | 9,587 | -### Pass (no retries) +#### Direct Success vs Failure (mean, abs & %) + +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 7,480 | - | - | +| GPT-4o-mini | 13,910 | - | - | +| DeepSeek-V3.1 | 13,760 | - | - | +| DeepSeek-R1 | 33,567 | - | - | +| Gemini-2.5 | 15,045 | 21,392 | +6,347 (42.2%) | +| Gemini-2.5-NT | 11,328 | 17,626 | +6,297 (55.6%) | +| Qwen3-235b | 12,176 | 9,587 | -2,589 (-21.3%) | + +### SocialMediaManager (aggregated across variants) + +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -953,9 +979,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 3 | 127,530 | -#### Pass (no retries) vs Failure (mean, abs & %) +#### Direct Success vs Failure (mean, abs & %) -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Failure mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 19,760 | 99,016 | +79,256 (401.1%) | | GPT-4o-mini | 36,934 | 219,309 | +182,375 (493.8%) | @@ -963,30 +989,4 @@ | DeepSeek-R1 | 42,398 | 227,416 | +185,018 (436.4%) | | Gemini-2.5 | 12,219 | - | - | | Gemini-2.5-NT | 9,820 | - | - | -| Qwen3-235b | 23,270 | 127,530 | +104,260 (448.0%) | - -### write_a_book_with_flows (aggregated across architectures) - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 180 | 163,074 | -| GPT-4o-mini | 180 | 125,190 | -| DeepSeek-V3.1 | 179 | 198,154 | -| DeepSeek-R1 | 163 | 156,596 | -| Gemini-2.5 | 180 | 102,955 | -| Gemini-2.5-NT | 179 | 78,117 | -| Qwen3-235b | 179 | 101,880 | - -#### Pass (no retries) vs Failure (mean, abs & %) - -| Model | Pass (no retries) mean | Failure mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 163,074 | - | - | -| GPT-4o-mini | 125,190 | - | - | -| DeepSeek-V3.1 | 198,154 | - | - | -| DeepSeek-R1 | 156,596 | - | - | -| Gemini-2.5 | 102,955 | - | - | -| Gemini-2.5-NT | 78,117 | - | - | -| Qwen3-235b | 101,880 | - | - | \ No newline at end of file +| Qwen3-235b | 23,270 | 127,530 | +104,260 (448.0%) | \ No newline at end of file diff --git a/data/processed/RQ3/Violin/project_token_stats_pass_vs_retry.md b/data/processed/RQ3/Violin/project_token_stats_pass_vs_retry.md index 446d4499958d7eec98c81e704dd5332fd7ad57d0..bcd4bde9567c7d5196d0e60b93cc1fd7e2bf613f 100644 --- a/data/processed/RQ3/Violin/project_token_stats_pass_vs_retry.md +++ b/data/processed/RQ3/Violin/project_token_stats_pass_vs_retry.md @@ -4,126 +4,164 @@ | Status | n | Mean | | --- | --- | --- | -| Pass (no retries) | 6849 | 47,691 | -| Pass (with retries) | 1179 | 83,133 | +| Direct Success | 6849 | 47,691 | +| Retry Success | 1179 | 83,133 | -## SQL_assistant-A2A +## BookWriter-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 36 | 8,736 | -| GPT-4o-mini | 50 | 17,092 | +| GPT-5 | 60 | 190,842 | +| GPT-4o-mini | 60 | 133,339 | +| DeepSeek-V3.1 | 60 | 184,315 | +| DeepSeek-R1 | 52 | 171,033 | +| Gemini-2.5 | 60 | 107,154 | +| Gemini-2.5-NT | 59 | 82,348 | +| Qwen3-235b | 59 | 110,871 | + +### Retry Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 55,457 | -| Gemini-2.5 | 2 | 23,118 | -| Gemini-2.5-NT | 2 | 36,636 | -| Qwen3-235b | 35 | 13,848 | +| DeepSeek-R1 | 8 | 318,572 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 1 | 94,226 | +| Qwen3-235b | 1 | 101,570 | + +### Direct Success vs Retry Success (mean, abs & %) + +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 190,842 | - | - | +| GPT-4o-mini | 133,339 | - | - | +| DeepSeek-V3.1 | 184,315 | - | - | +| DeepSeek-R1 | 171,033 | 318,572 | +147,539 (86.3%) | +| Gemini-2.5 | 107,154 | - | - | +| Gemini-2.5-NT | 82,348 | 94,226 | +11,878 (14.4%) | +| Qwen3-235b | 110,871 | 101,570 | -9,301 (-8.4%) | + +## BookWriter-A2A -### Pass (with retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 24 | 14,869 | -| GPT-4o-mini | 10 | 26,350 | -| DeepSeek-V3.1 | 60 | 36,805 | -| DeepSeek-R1 | 57 | 98,627 | -| Gemini-2.5 | 20 | 44,005 | -| Gemini-2.5-NT | 5 | 71,593 | -| Qwen3-235b | 23 | 23,436 | +| GPT-5 | 60 | 181,022 | +| GPT-4o-mini | 60 | 139,665 | +| DeepSeek-V3.1 | 59 | 270,142 | +| DeepSeek-R1 | 52 | 173,528 | +| Gemini-2.5 | 60 | 110,126 | +| Gemini-2.5-NT | 60 | 86,249 | +| Qwen3-235b | 60 | 118,406 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Retry Success -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 1 | 398,139 | +| DeepSeek-R1 | 8 | 272,152 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 0 | - | + +### Direct Success vs Retry Success (mean, abs & %) + +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 8,736 | 14,869 | +6,133 (70.2%) | -| GPT-4o-mini | 17,092 | 26,350 | +9,258 (54.2%) | -| DeepSeek-V3.1 | - | 36,805 | - | -| DeepSeek-R1 | 55,457 | 98,627 | +43,170 (77.8%) | -| Gemini-2.5 | 23,118 | 44,005 | +20,887 (90.4%) | -| Gemini-2.5-NT | 36,636 | 71,593 | +34,956 (95.4%) | -| Qwen3-235b | 13,848 | 23,436 | +9,588 (69.2%) | +| GPT-5 | 181,022 | - | - | +| GPT-4o-mini | 139,665 | - | - | +| DeepSeek-V3.1 | 270,142 | 398,139 | +127,997 (47.4%) | +| DeepSeek-R1 | 173,528 | 272,152 | +98,623 (56.8%) | +| Gemini-2.5 | 110,126 | - | - | +| Gemini-2.5-NT | 86,249 | - | - | +| Qwen3-235b | 118,406 | - | - | -## SQL_assistant-A2A_mix +## BookWriter-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 46 | 6,953 | -| GPT-4o-mini | 43 | 10,004 | -| DeepSeek-V3.1 | 44 | 12,571 | -| DeepSeek-R1 | 20 | 29,145 | -| Gemini-2.5 | 25 | 14,332 | -| Gemini-2.5-NT | 33 | 9,561 | -| Qwen3-235b | 40 | 10,594 | +| GPT-5 | 60 | 117,358 | +| GPT-4o-mini | 60 | 102,564 | +| DeepSeek-V3.1 | 60 | 141,204 | +| DeepSeek-R1 | 59 | 128,948 | +| Gemini-2.5 | 60 | 91,585 | +| Gemini-2.5-NT | 60 | 65,826 | +| Qwen3-235b | 60 | 76,512 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 14 | 15,537 | -| GPT-4o-mini | 17 | 23,168 | -| DeepSeek-V3.1 | 16 | 27,226 | -| DeepSeek-R1 | 40 | 89,473 | -| Gemini-2.5 | 26 | 30,481 | -| Gemini-2.5-NT | 19 | 33,544 | -| Qwen3-235b | 20 | 23,149 | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 1 | 185,051 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 0 | - | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 6,953 | 15,537 | +8,584 (123.5%) | -| GPT-4o-mini | 10,004 | 23,168 | +13,165 (131.6%) | -| DeepSeek-V3.1 | 12,571 | 27,226 | +14,655 (116.6%) | -| DeepSeek-R1 | 29,145 | 89,473 | +60,327 (207.0%) | -| Gemini-2.5 | 14,332 | 30,481 | +16,149 (112.7%) | -| Gemini-2.5-NT | 9,561 | 33,544 | +23,983 (250.8%) | -| Qwen3-235b | 10,594 | 23,149 | +12,555 (118.5%) | +| GPT-5 | 117,358 | - | - | +| GPT-4o-mini | 102,564 | - | - | +| DeepSeek-V3.1 | 141,204 | - | - | +| DeepSeek-R1 | 128,948 | 185,051 | +56,103 (43.5%) | +| Gemini-2.5 | 91,585 | - | - | +| Gemini-2.5-NT | 65,826 | - | - | +| Qwen3-235b | 76,512 | - | - | -## SQL_assistant-MCP +## EmailResponder -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 19 | 6,378 | -| GPT-4o-mini | 21 | 14,334 | -| DeepSeek-V3.1 | 26 | 15,771 | -| DeepSeek-R1 | 13 | 35,319 | -| Gemini-2.5 | 6 | 15,321 | -| Gemini-2.5-NT | 10 | 12,100 | -| Qwen3-235b | 37 | 12,304 | +| GPT-5 | 100 | 5,562 | +| GPT-4o-mini | 100 | 8,044 | +| DeepSeek-V3.1 | 98 | 7,350 | +| DeepSeek-R1 | 72 | 13,490 | +| Gemini-2.5 | 90 | 11,211 | +| Gemini-2.5-NT | 88 | 12,494 | +| Qwen3-235b | 100 | 5,900 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 41 | 17,358 | -| GPT-4o-mini | 39 | 29,576 | -| DeepSeek-V3.1 | 34 | 33,181 | -| DeepSeek-R1 | 47 | 93,025 | -| Gemini-2.5 | 11 | 42,818 | -| Gemini-2.5-NT | 6 | 33,419 | -| Qwen3-235b | 23 | 25,159 | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 2 | 11,414 | +| DeepSeek-R1 | 26 | 26,688 | +| Gemini-2.5 | 9 | 233,862 | +| Gemini-2.5-NT | 12 | 261,883 | +| Qwen3-235b | 0 | - | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 6,378 | 17,358 | +10,981 (172.2%) | -| GPT-4o-mini | 14,334 | 29,576 | +15,242 (106.3%) | -| DeepSeek-V3.1 | 15,771 | 33,181 | +17,409 (110.4%) | -| DeepSeek-R1 | 35,319 | 93,025 | +57,706 (163.4%) | -| Gemini-2.5 | 15,321 | 42,818 | +27,497 (179.5%) | -| Gemini-2.5-NT | 12,100 | 33,419 | +21,319 (176.2%) | -| Qwen3-235b | 12,304 | 25,159 | +12,855 (104.5%) | +| GPT-5 | 5,562 | - | - | +| GPT-4o-mini | 8,044 | - | - | +| DeepSeek-V3.1 | 7,350 | 11,414 | +4,064 (55.3%) | +| DeepSeek-R1 | 13,490 | 26,688 | +13,198 (97.8%) | +| Gemini-2.5 | 11,211 | 233,862 | +222,651 (1986.0%) | +| Gemini-2.5-NT | 12,494 | 261,883 | +249,388 (1996.0%) | +| Qwen3-235b | 5,900 | - | - | -## email_auto_responder_flow-MCP +## EmailResponder-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -135,7 +173,7 @@ | Gemini-2.5-NT | 88 | 12,696 | | Qwen3-235b | 100 | 7,434 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -147,9 +185,9 @@ | Gemini-2.5-NT | 12 | 199,673 | | Qwen3-235b | 0 | - | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 6,903 | - | - | | GPT-4o-mini | 9,437 | - | - | @@ -159,47 +197,47 @@ | Gemini-2.5-NT | 12,696 | 199,673 | +186,977 (1472.7%) | | Qwen3-235b | 7,434 | - | - | -## email_auto_responder_flow +## GameBuilder -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 100 | 5,562 | -| GPT-4o-mini | 100 | 8,044 | -| DeepSeek-V3.1 | 98 | 7,350 | -| DeepSeek-R1 | 72 | 13,490 | -| Gemini-2.5 | 90 | 11,211 | -| Gemini-2.5-NT | 88 | 12,494 | -| Qwen3-235b | 100 | 5,900 | +| GPT-5 | 25 | 22,904 | +| GPT-4o-mini | 25 | 17,823 | +| DeepSeek-V3.1 | 24 | 39,841 | +| DeepSeek-R1 | 25 | 44,536 | +| Gemini-2.5 | 25 | 72,772 | +| Gemini-2.5-NT | 23 | 54,423 | +| Qwen3-235b | 25 | 24,212 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 2 | 11,414 | -| DeepSeek-R1 | 26 | 26,688 | -| Gemini-2.5 | 9 | 233,862 | -| Gemini-2.5-NT | 12 | 261,883 | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 2 | 223,682 | | Qwen3-235b | 0 | - | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 5,562 | - | - | -| GPT-4o-mini | 8,044 | - | - | -| DeepSeek-V3.1 | 7,350 | 11,414 | +4,064 (55.3%) | -| DeepSeek-R1 | 13,490 | 26,688 | +13,198 (97.8%) | -| Gemini-2.5 | 11,211 | 233,862 | +222,651 (1986.0%) | -| Gemini-2.5-NT | 12,494 | 261,883 | +249,388 (1996.0%) | -| Qwen3-235b | 5,900 | - | - | +| GPT-5 | 22,904 | - | - | +| GPT-4o-mini | 17,823 | - | - | +| DeepSeek-V3.1 | 39,841 | - | - | +| DeepSeek-R1 | 44,536 | - | - | +| Gemini-2.5 | 72,772 | - | - | +| Gemini-2.5-NT | 54,423 | 223,682 | +169,258 (311.0%) | +| Qwen3-235b | 24,212 | - | - | -## game_builder-MCP +## GameBuilder-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -211,9 +249,9 @@ | Gemini-2.5-NT | 25 | 50,782 | | Qwen3-235b | 25 | 24,039 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 22,425 | - | - | | GPT-4o-mini | 47,996 | - | - | @@ -223,187 +261,73 @@ | Gemini-2.5-NT | 50,782 | - | - | | Qwen3-235b | 24,039 | - | - | -## game_builder +## LandingPageGenerator-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 25 | 22,904 | -| GPT-4o-mini | 25 | 17,823 | -| DeepSeek-V3.1 | 24 | 39,841 | -| DeepSeek-R1 | 25 | 44,536 | -| Gemini-2.5 | 25 | 72,772 | -| Gemini-2.5-NT | 23 | 54,423 | -| Qwen3-235b | 25 | 24,212 | +| GPT-5 | 60 | 12,653 | +| GPT-4o-mini | 59 | 19,333 | +| DeepSeek-V3.1 | 55 | 15,912 | +| DeepSeek-R1 | 33 | 24,081 | +| Gemini-2.5 | 1 | 55,158 | +| Gemini-2.5-NT | 2 | 188,992 | +| Qwen3-235b | 53 | 15,068 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | +| DeepSeek-R1 | 8 | 61,223 | | Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 2 | 223,682 | +| Gemini-2.5-NT | 0 | - | | Qwen3-235b | 0 | - | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 22,904 | - | - | -| GPT-4o-mini | 17,823 | - | - | -| DeepSeek-V3.1 | 39,841 | - | - | -| DeepSeek-R1 | 44,536 | - | - | -| Gemini-2.5 | 72,772 | - | - | -| Gemini-2.5-NT | 54,423 | 223,682 | +169,258 (311.0%) | -| Qwen3-235b | 24,212 | - | - | +| GPT-5 | 12,653 | - | - | +| GPT-4o-mini | 19,333 | - | - | +| DeepSeek-V3.1 | 15,912 | - | - | +| DeepSeek-R1 | 24,081 | 61,223 | +37,142 (154.2%) | +| Gemini-2.5 | 55,158 | - | - | +| Gemini-2.5-NT | 188,992 | - | - | +| Qwen3-235b | 15,068 | - | - | -## intelligent_recruitment_platform-A2A +## LandingPageGenerator-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 31,991 | -| GPT-4o-mini | 60 | 74,951 | -| DeepSeek-V3.1 | 60 | 65,662 | -| DeepSeek-R1 | 56 | 73,964 | -| Gemini-2.5 | 60 | 68,719 | -| Gemini-2.5-NT | 60 | 55,660 | -| Qwen3-235b | 44 | 73,316 | +| GPT-5 | 16 | 18,863 | +| GPT-4o-mini | 58 | 20,076 | +| DeepSeek-V3.1 | 60 | 28,814 | +| DeepSeek-R1 | 7 | 46,577 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 27 | 20,066 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 68,009 | +| DeepSeek-R1 | 0 | - | | Gemini-2.5 | 0 | - | | Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 15 | 164,476 | +| Qwen3-235b | 3 | 34,203 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 31,991 | - | - | -| GPT-4o-mini | 74,951 | - | - | -| DeepSeek-V3.1 | 65,662 | - | - | -| DeepSeek-R1 | 73,964 | 68,009 | -5,956 (-8.1%) | -| Gemini-2.5 | 68,719 | - | - | -| Gemini-2.5-NT | 55,660 | - | - | -| Qwen3-235b | 73,316 | 164,476 | +91,160 (124.3%) | - -## intelligent_recruitment_platform-A2A_mix - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 32,672 | -| GPT-4o-mini | 60 | 101,936 | -| DeepSeek-V3.1 | 60 | 65,048 | -| DeepSeek-R1 | 57 | 76,741 | -| Gemini-2.5 | 60 | 57,128 | -| Gemini-2.5-NT | 60 | 36,887 | -| Qwen3-235b | 49 | 67,089 | - -### Pass (with retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 3 | 122,480 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 11 | 100,743 | - -### Pass (no retries) vs Pass (with retries) (mean, abs & %) - -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 32,672 | - | - | -| GPT-4o-mini | 101,936 | - | - | -| DeepSeek-V3.1 | 65,048 | - | - | -| DeepSeek-R1 | 76,741 | 122,480 | +45,739 (59.6%) | -| Gemini-2.5 | 57,128 | - | - | -| Gemini-2.5-NT | 36,887 | - | - | -| Qwen3-235b | 67,089 | 100,743 | +33,653 (50.2%) | - -## intelligent_recruitment_platform-MCP - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 32,411 | -| GPT-4o-mini | 60 | 99,933 | -| DeepSeek-V3.1 | 60 | 65,748 | -| DeepSeek-R1 | 58 | 82,498 | -| Gemini-2.5 | 60 | 64,864 | -| Gemini-2.5-NT | 60 | 54,402 | -| Qwen3-235b | 40 | 68,264 | - -### Pass (with retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 2 | 148,137 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 18 | 124,320 | - -### Pass (no retries) vs Pass (with retries) (mean, abs & %) - -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 32,411 | - | - | -| GPT-4o-mini | 99,933 | - | - | -| DeepSeek-V3.1 | 65,748 | - | - | -| DeepSeek-R1 | 82,498 | 148,137 | +65,639 (79.6%) | -| Gemini-2.5 | 64,864 | - | - | -| Gemini-2.5-NT | 54,402 | - | - | -| Qwen3-235b | 68,264 | 124,320 | +56,055 (82.1%) | - -## landing_page_generator-A2A - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 16 | 18,863 | -| GPT-4o-mini | 58 | 20,076 | -| DeepSeek-V3.1 | 60 | 28,814 | -| DeepSeek-R1 | 7 | 46,577 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 27 | 20,066 | - -### Pass (with retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 3 | 34,203 | - -### Pass (no retries) vs Pass (with retries) (mean, abs & %) - -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 18,863 | - | - | | GPT-4o-mini | 20,076 | - | - | @@ -413,9 +337,9 @@ | Gemini-2.5-NT | - | - | - | | Qwen3-235b | 20,066 | 34,203 | +14,137 (70.5%) | -## landing_page_generator-A2A_mix +## LandingPageGenerator-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -427,7 +351,7 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 29 | 19,428 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -439,9 +363,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 2 | 33,284 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 29,492 | - | - | | GPT-4o-mini | 43,589 | - | - | @@ -451,47 +375,47 @@ | Gemini-2.5-NT | - | - | - | | Qwen3-235b | 19,428 | 33,284 | +13,856 (71.3%) | -## landing_page_generator-MCP +## MarkdownValidator -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 12,653 | -| GPT-4o-mini | 59 | 19,333 | -| DeepSeek-V3.1 | 55 | 15,912 | -| DeepSeek-R1 | 33 | 24,081 | -| Gemini-2.5 | 1 | 55,158 | -| Gemini-2.5-NT | 2 | 188,992 | -| Qwen3-235b | 53 | 15,068 | +| GPT-5 | 60 | 1,853 | +| GPT-4o-mini | 60 | 1,853 | +| DeepSeek-V3.1 | 60 | 1,948 | +| DeepSeek-R1 | 60 | 5,912 | +| Gemini-2.5 | 47 | 81,437 | +| Gemini-2.5-NT | 57 | 16,246 | +| Qwen3-235b | 60 | 1,854 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 8 | 61,223 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | +| DeepSeek-R1 | 0 | - | +| Gemini-2.5 | 13 | 329,685 | +| Gemini-2.5-NT | 3 | 317,934 | | Qwen3-235b | 0 | - | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 12,653 | - | - | -| GPT-4o-mini | 19,333 | - | - | -| DeepSeek-V3.1 | 15,912 | - | - | -| DeepSeek-R1 | 24,081 | 61,223 | +37,142 (154.2%) | -| Gemini-2.5 | 55,158 | - | - | -| Gemini-2.5-NT | 188,992 | - | - | -| Qwen3-235b | 15,068 | - | - | +| GPT-5 | 1,853 | - | - | +| GPT-4o-mini | 1,853 | - | - | +| DeepSeek-V3.1 | 1,948 | - | - | +| DeepSeek-R1 | 5,912 | - | - | +| Gemini-2.5 | 81,437 | 329,685 | +248,248 (304.8%) | +| Gemini-2.5-NT | 16,246 | 317,934 | +301,688 (1857.0%) | +| Qwen3-235b | 1,854 | - | - | -## markdown_validator-MCP +## MarkdownValidator-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -503,7 +427,7 @@ | Gemini-2.5-NT | 56 | 39,448 | | Qwen3-235b | 60 | 1,994 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -515,9 +439,9 @@ | Gemini-2.5-NT | 4 | 376,206 | | Qwen3-235b | 0 | - | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 1,967 | - | - | | GPT-4o-mini | 1,942 | - | - | @@ -527,123 +451,237 @@ | Gemini-2.5-NT | 39,448 | 376,206 | +336,758 (853.7%) | | Qwen3-235b | 1,994 | - | - | -## markdown_validator +## RecruitmentAssistant-MCP -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 1,853 | -| GPT-4o-mini | 60 | 1,853 | -| DeepSeek-V3.1 | 60 | 1,948 | -| DeepSeek-R1 | 60 | 5,912 | -| Gemini-2.5 | 47 | 81,437 | -| Gemini-2.5-NT | 57 | 16,246 | -| Qwen3-235b | 60 | 1,854 | +| GPT-5 | 60 | 32,411 | +| GPT-4o-mini | 60 | 99,933 | +| DeepSeek-V3.1 | 60 | 65,748 | +| DeepSeek-R1 | 58 | 82,498 | +| Gemini-2.5 | 60 | 64,864 | +| Gemini-2.5-NT | 60 | 54,402 | +| Qwen3-235b | 40 | 68,264 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | | GPT-5 | 0 | - | | GPT-4o-mini | 0 | - | | DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 0 | - | -| Gemini-2.5 | 13 | 329,685 | -| Gemini-2.5-NT | 3 | 317,934 | -| Qwen3-235b | 0 | - | +| DeepSeek-R1 | 2 | 148,137 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 18 | 124,320 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 1,853 | - | - | -| GPT-4o-mini | 1,853 | - | - | -| DeepSeek-V3.1 | 1,948 | - | - | -| DeepSeek-R1 | 5,912 | - | - | -| Gemini-2.5 | 81,437 | 329,685 | +248,248 (304.8%) | -| Gemini-2.5-NT | 16,246 | 317,934 | +301,688 (1857.0%) | -| Qwen3-235b | 1,854 | - | - | +| GPT-5 | 32,411 | - | - | +| GPT-4o-mini | 99,933 | - | - | +| DeepSeek-V3.1 | 65,748 | - | - | +| DeepSeek-R1 | 82,498 | 148,137 | +65,639 (79.6%) | +| Gemini-2.5 | 64,864 | - | - | +| Gemini-2.5-NT | 54,402 | - | - | +| Qwen3-235b | 68,264 | 124,320 | +56,055 (82.1%) | -## self_evaluation_loop_flow-A2A +## RecruitmentAssistant-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 31 | 23,996 | -| GPT-4o-mini | 24 | 42,258 | -| DeepSeek-V3.1 | 46 | 70,407 | -| DeepSeek-R1 | 21 | 45,776 | -| Gemini-2.5 | 56 | 12,176 | -| Gemini-2.5-NT | 53 | 10,675 | -| Qwen3-235b | 27 | 30,147 | +| GPT-5 | 60 | 31,991 | +| GPT-4o-mini | 60 | 74,951 | +| DeepSeek-V3.1 | 60 | 65,662 | +| DeepSeek-R1 | 56 | 73,964 | +| Gemini-2.5 | 60 | 68,719 | +| Gemini-2.5-NT | 60 | 55,660 | +| Qwen3-235b | 44 | 73,316 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 27 | 55,285 | -| GPT-4o-mini | 32 | 125,026 | -| DeepSeek-V3.1 | 14 | 143,090 | -| DeepSeek-R1 | 34 | 165,493 | -| Gemini-2.5 | 4 | 23,860 | -| Gemini-2.5-NT | 7 | 17,136 | -| Qwen3-235b | 32 | 72,094 | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 3 | 68,009 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 15 | 164,476 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 23,996 | 55,285 | +31,289 (130.4%) | -| GPT-4o-mini | 42,258 | 125,026 | +82,768 (195.9%) | -| DeepSeek-V3.1 | 70,407 | 143,090 | +72,683 (103.2%) | -| DeepSeek-R1 | 45,776 | 165,493 | +119,718 (261.5%) | -| Gemini-2.5 | 12,176 | 23,860 | +11,684 (96.0%) | -| Gemini-2.5-NT | 10,675 | 17,136 | +6,461 (60.5%) | -| Qwen3-235b | 30,147 | 72,094 | +41,947 (139.1%) | +| GPT-5 | 31,991 | - | - | +| GPT-4o-mini | 74,951 | - | - | +| DeepSeek-V3.1 | 65,662 | - | - | +| DeepSeek-R1 | 73,964 | 68,009 | -5,956 (-8.1%) | +| Gemini-2.5 | 68,719 | - | - | +| Gemini-2.5-NT | 55,660 | - | - | +| Qwen3-235b | 73,316 | 164,476 | +91,160 (124.3%) | -## self_evaluation_loop_flow-A2A_mix +## RecruitmentAssistant-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 55 | 15,857 | -| GPT-4o-mini | 51 | 33,008 | -| DeepSeek-V3.1 | 49 | 51,074 | -| DeepSeek-R1 | 27 | 40,199 | -| Gemini-2.5 | 35 | 13,495 | -| Gemini-2.5-NT | 45 | 7,487 | -| Qwen3-235b | 35 | 14,019 | +| GPT-5 | 60 | 32,672 | +| GPT-4o-mini | 60 | 101,936 | +| DeepSeek-V3.1 | 60 | 65,048 | +| DeepSeek-R1 | 57 | 76,741 | +| Gemini-2.5 | 60 | 57,128 | +| Gemini-2.5-NT | 60 | 36,887 | +| Qwen3-235b | 49 | 67,089 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 5 | 27,133 | -| GPT-4o-mini | 9 | 101,766 | -| DeepSeek-V3.1 | 11 | 85,846 | -| DeepSeek-R1 | 31 | 133,368 | -| Gemini-2.5 | 25 | 30,123 | -| Gemini-2.5-NT | 15 | 16,107 | -| Qwen3-235b | 24 | 51,432 | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 3 | 122,480 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 11 | 100,743 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 15,857 | 27,133 | +11,276 (71.1%) | -| GPT-4o-mini | 33,008 | 101,766 | +68,758 (208.3%) | -| DeepSeek-V3.1 | 51,074 | 85,846 | +34,772 (68.1%) | -| DeepSeek-R1 | 40,199 | 133,368 | +93,169 (231.8%) | -| Gemini-2.5 | 13,495 | 30,123 | +16,628 (123.2%) | -| Gemini-2.5-NT | 7,487 | 16,107 | +8,620 (115.1%) | -| Qwen3-235b | 14,019 | 51,432 | +37,413 (266.9%) | +| GPT-5 | 32,672 | - | - | +| GPT-4o-mini | 101,936 | - | - | +| DeepSeek-V3.1 | 65,048 | - | - | +| DeepSeek-R1 | 76,741 | 122,480 | +45,739 (59.6%) | +| Gemini-2.5 | 57,128 | - | - | +| Gemini-2.5-NT | 36,887 | - | - | +| Qwen3-235b | 67,089 | 100,743 | +33,653 (50.2%) | + +## SQLAssistant-MCP + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 19 | 6,378 | +| GPT-4o-mini | 21 | 14,334 | +| DeepSeek-V3.1 | 26 | 15,771 | +| DeepSeek-R1 | 13 | 35,319 | +| Gemini-2.5 | 6 | 15,321 | +| Gemini-2.5-NT | 10 | 12,100 | +| Qwen3-235b | 37 | 12,304 | + +### Retry Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 41 | 17,358 | +| GPT-4o-mini | 39 | 29,576 | +| DeepSeek-V3.1 | 34 | 33,181 | +| DeepSeek-R1 | 47 | 93,025 | +| Gemini-2.5 | 11 | 42,818 | +| Gemini-2.5-NT | 6 | 33,419 | +| Qwen3-235b | 23 | 25,159 | + +### Direct Success vs Retry Success (mean, abs & %) + +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 6,378 | 17,358 | +10,981 (172.2%) | +| GPT-4o-mini | 14,334 | 29,576 | +15,242 (106.3%) | +| DeepSeek-V3.1 | 15,771 | 33,181 | +17,409 (110.4%) | +| DeepSeek-R1 | 35,319 | 93,025 | +57,706 (163.4%) | +| Gemini-2.5 | 15,321 | 42,818 | +27,497 (179.5%) | +| Gemini-2.5-NT | 12,100 | 33,419 | +21,319 (176.2%) | +| Qwen3-235b | 12,304 | 25,159 | +12,855 (104.5%) | + +## SQLAssistant-A2A + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 36 | 8,736 | +| GPT-4o-mini | 50 | 17,092 | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 3 | 55,457 | +| Gemini-2.5 | 2 | 23,118 | +| Gemini-2.5-NT | 2 | 36,636 | +| Qwen3-235b | 35 | 13,848 | -## self_evaluation_loop_flow-MCP +### Retry Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 24 | 14,869 | +| GPT-4o-mini | 10 | 26,350 | +| DeepSeek-V3.1 | 60 | 36,805 | +| DeepSeek-R1 | 57 | 98,627 | +| Gemini-2.5 | 20 | 44,005 | +| Gemini-2.5-NT | 5 | 71,593 | +| Qwen3-235b | 23 | 23,436 | -### Pass (no retries) +### Direct Success vs Retry Success (mean, abs & %) + +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 8,736 | 14,869 | +6,133 (70.2%) | +| GPT-4o-mini | 17,092 | 26,350 | +9,258 (54.2%) | +| DeepSeek-V3.1 | - | 36,805 | - | +| DeepSeek-R1 | 55,457 | 98,627 | +43,170 (77.8%) | +| Gemini-2.5 | 23,118 | 44,005 | +20,887 (90.4%) | +| Gemini-2.5-NT | 36,636 | 71,593 | +34,956 (95.4%) | +| Qwen3-235b | 13,848 | 23,436 | +9,588 (69.2%) | + +## SQLAssistant-H-A2A + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 46 | 6,953 | +| GPT-4o-mini | 43 | 10,004 | +| DeepSeek-V3.1 | 44 | 12,571 | +| DeepSeek-R1 | 20 | 29,145 | +| Gemini-2.5 | 25 | 14,332 | +| Gemini-2.5-NT | 33 | 9,561 | +| Qwen3-235b | 40 | 10,594 | + +### Retry Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 14 | 15,537 | +| GPT-4o-mini | 17 | 23,168 | +| DeepSeek-V3.1 | 16 | 27,226 | +| DeepSeek-R1 | 40 | 89,473 | +| Gemini-2.5 | 26 | 30,481 | +| Gemini-2.5-NT | 19 | 33,544 | +| Qwen3-235b | 20 | 23,149 | + +### Direct Success vs Retry Success (mean, abs & %) + +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 6,953 | 15,537 | +8,584 (123.5%) | +| GPT-4o-mini | 10,004 | 23,168 | +13,165 (131.6%) | +| DeepSeek-V3.1 | 12,571 | 27,226 | +14,655 (116.6%) | +| DeepSeek-R1 | 29,145 | 89,473 | +60,327 (207.0%) | +| Gemini-2.5 | 14,332 | 30,481 | +16,149 (112.7%) | +| Gemini-2.5-NT | 9,561 | 33,544 | +23,983 (250.8%) | +| Qwen3-235b | 10,594 | 23,149 | +12,555 (118.5%) | + +## SocialMediaManager-MCP + +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -655,7 +693,7 @@ | Gemini-2.5-NT | 55 | 10,904 | | Qwen3-235b | 35 | 27,216 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -667,9 +705,9 @@ | Gemini-2.5-NT | 5 | 17,821 | | Qwen3-235b | 24 | 72,110 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 22,287 | 57,749 | +35,462 (159.1%) | | GPT-4o-mini | 40,555 | 134,943 | +94,388 (232.7%) | @@ -679,163 +717,125 @@ | Gemini-2.5-NT | 10,904 | 17,821 | +6,917 (63.4%) | | Qwen3-235b | 27,216 | 72,110 | +44,894 (165.0%) | -## write_a_book_with_flows-A2A - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 60 | 181,022 | -| GPT-4o-mini | 60 | 139,665 | -| DeepSeek-V3.1 | 59 | 270,142 | -| DeepSeek-R1 | 52 | 173,528 | -| Gemini-2.5 | 60 | 110,126 | -| Gemini-2.5-NT | 60 | 86,249 | -| Qwen3-235b | 60 | 118,406 | - -### Pass (with retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 1 | 398,139 | -| DeepSeek-R1 | 8 | 272,152 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 0 | - | - -### Pass (no retries) vs Pass (with retries) (mean, abs & %) - -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 181,022 | - | - | -| GPT-4o-mini | 139,665 | - | - | -| DeepSeek-V3.1 | 270,142 | 398,139 | +127,997 (47.4%) | -| DeepSeek-R1 | 173,528 | 272,152 | +98,623 (56.8%) | -| Gemini-2.5 | 110,126 | - | - | -| Gemini-2.5-NT | 86,249 | - | - | -| Qwen3-235b | 118,406 | - | - | - -## write_a_book_with_flows-A2A_mix +## SocialMediaManager-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 117,358 | -| GPT-4o-mini | 60 | 102,564 | -| DeepSeek-V3.1 | 60 | 141,204 | -| DeepSeek-R1 | 59 | 128,948 | -| Gemini-2.5 | 60 | 91,585 | -| Gemini-2.5-NT | 60 | 65,826 | -| Qwen3-235b | 60 | 76,512 | +| GPT-5 | 31 | 23,996 | +| GPT-4o-mini | 24 | 42,258 | +| DeepSeek-V3.1 | 46 | 70,407 | +| DeepSeek-R1 | 21 | 45,776 | +| Gemini-2.5 | 56 | 12,176 | +| Gemini-2.5-NT | 53 | 10,675 | +| Qwen3-235b | 27 | 30,147 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 1 | 185,051 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 0 | - | +| GPT-5 | 27 | 55,285 | +| GPT-4o-mini | 32 | 125,026 | +| DeepSeek-V3.1 | 14 | 143,090 | +| DeepSeek-R1 | 34 | 165,493 | +| Gemini-2.5 | 4 | 23,860 | +| Gemini-2.5-NT | 7 | 17,136 | +| Qwen3-235b | 32 | 72,094 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 117,358 | - | - | -| GPT-4o-mini | 102,564 | - | - | -| DeepSeek-V3.1 | 141,204 | - | - | -| DeepSeek-R1 | 128,948 | 185,051 | +56,103 (43.5%) | -| Gemini-2.5 | 91,585 | - | - | -| Gemini-2.5-NT | 65,826 | - | - | -| Qwen3-235b | 76,512 | - | - | +| GPT-5 | 23,996 | 55,285 | +31,289 (130.4%) | +| GPT-4o-mini | 42,258 | 125,026 | +82,768 (195.9%) | +| DeepSeek-V3.1 | 70,407 | 143,090 | +72,683 (103.2%) | +| DeepSeek-R1 | 45,776 | 165,493 | +119,718 (261.5%) | +| Gemini-2.5 | 12,176 | 23,860 | +11,684 (96.0%) | +| Gemini-2.5-NT | 10,675 | 17,136 | +6,461 (60.5%) | +| Qwen3-235b | 30,147 | 72,094 | +41,947 (139.1%) | -## write_a_book_with_flows-MCP +## SocialMediaManager-H-A2A -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 60 | 190,842 | -| GPT-4o-mini | 60 | 133,339 | -| DeepSeek-V3.1 | 60 | 184,315 | -| DeepSeek-R1 | 52 | 171,033 | -| Gemini-2.5 | 60 | 107,154 | -| Gemini-2.5-NT | 59 | 82,348 | -| Qwen3-235b | 59 | 110,871 | +| GPT-5 | 55 | 15,857 | +| GPT-4o-mini | 51 | 33,008 | +| DeepSeek-V3.1 | 49 | 51,074 | +| DeepSeek-R1 | 27 | 40,199 | +| Gemini-2.5 | 35 | 13,495 | +| Gemini-2.5-NT | 45 | 7,487 | +| Qwen3-235b | 35 | 14,019 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 8 | 318,572 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 1 | 94,226 | -| Qwen3-235b | 1 | 101,570 | +| GPT-5 | 5 | 27,133 | +| GPT-4o-mini | 9 | 101,766 | +| DeepSeek-V3.1 | 11 | 85,846 | +| DeepSeek-R1 | 31 | 133,368 | +| Gemini-2.5 | 25 | 30,123 | +| Gemini-2.5-NT | 15 | 16,107 | +| Qwen3-235b | 24 | 51,432 | -### Pass (no retries) vs Pass (with retries) (mean, abs & %) +### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 190,842 | - | - | -| GPT-4o-mini | 133,339 | - | - | -| DeepSeek-V3.1 | 184,315 | - | - | -| DeepSeek-R1 | 171,033 | 318,572 | +147,539 (86.3%) | -| Gemini-2.5 | 107,154 | - | - | -| Gemini-2.5-NT | 82,348 | 94,226 | +11,878 (14.4%) | -| Qwen3-235b | 110,871 | 101,570 | -9,301 (-8.4%) | +| GPT-5 | 15,857 | 27,133 | +11,276 (71.1%) | +| GPT-4o-mini | 33,008 | 101,766 | +68,758 (208.3%) | +| DeepSeek-V3.1 | 51,074 | 85,846 | +34,772 (68.1%) | +| DeepSeek-R1 | 40,199 | 133,368 | +93,169 (231.8%) | +| Gemini-2.5 | 13,495 | 30,123 | +16,628 (123.2%) | +| Gemini-2.5-NT | 7,487 | 16,107 | +8,620 (115.1%) | +| Qwen3-235b | 14,019 | 51,432 | +37,413 (266.9%) | ## Series aggregates (by task prefix) -### SQL_assistant (aggregated across architectures) +### BookWriter (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 101 | 7,480 | -| GPT-4o-mini | 114 | 13,910 | -| DeepSeek-V3.1 | 70 | 13,760 | -| DeepSeek-R1 | 36 | 33,567 | -| Gemini-2.5 | 33 | 15,045 | -| Gemini-2.5-NT | 45 | 11,328 | -| Qwen3-235b | 112 | 12,176 | +| GPT-5 | 180 | 163,074 | +| GPT-4o-mini | 180 | 125,190 | +| DeepSeek-V3.1 | 179 | 198,154 | +| DeepSeek-R1 | 163 | 156,596 | +| Gemini-2.5 | 180 | 102,955 | +| Gemini-2.5-NT | 179 | 78,117 | +| Qwen3-235b | 179 | 101,880 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | -| GPT-5 | 79 | 16,279 | -| GPT-4o-mini | 66 | 27,437 | -| DeepSeek-V3.1 | 110 | 34,291 | -| DeepSeek-R1 | 144 | 94,256 | -| Gemini-2.5 | 57 | 37,607 | -| Gemini-2.5-NT | 30 | 39,861 | -| Qwen3-235b | 66 | 23,950 | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 1 | 398,139 | +| DeepSeek-R1 | 17 | 288,873 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 1 | 94,226 | +| Qwen3-235b | 1 | 101,570 | -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) +#### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | -| GPT-5 | 7,480 | 16,279 | +8,799 (117.6%) | -| GPT-4o-mini | 13,910 | 27,437 | +13,527 (97.2%) | -| DeepSeek-V3.1 | 13,760 | 34,291 | +20,532 (149.2%) | -| DeepSeek-R1 | 33,567 | 94,256 | +60,689 (180.8%) | -| Gemini-2.5 | 15,045 | 37,607 | +22,563 (150.0%) | -| Gemini-2.5-NT | 11,328 | 39,861 | +28,532 (251.9%) | -| Qwen3-235b | 12,176 | 23,950 | +11,774 (96.7%) | +| GPT-5 | 163,074 | - | - | +| GPT-4o-mini | 125,190 | - | - | +| DeepSeek-V3.1 | 198,154 | 398,139 | +199,985 (100.9%) | +| DeepSeek-R1 | 156,596 | 288,873 | +132,277 (84.5%) | +| Gemini-2.5 | 102,955 | - | - | +| Gemini-2.5-NT | 78,117 | 94,226 | +16,109 (20.6%) | +| Qwen3-235b | 101,880 | 101,570 | -310 (-0.3%) | -### email_auto_responder_flow (aggregated across architectures) +### EmailResponder (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -847,7 +847,7 @@ | Gemini-2.5-NT | 176 | 12,595 | | Qwen3-235b | 200 | 6,667 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -859,9 +859,9 @@ | Gemini-2.5-NT | 24 | 230,778 | | Qwen3-235b | 0 | - | -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) +#### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 6,232 | - | - | | GPT-4o-mini | 8,741 | - | - | @@ -871,9 +871,9 @@ | Gemini-2.5-NT | 12,595 | 230,778 | +218,183 (1732.3%) | | Qwen3-235b | 6,667 | - | - | -### game_builder (aggregated across architectures) +### GameBuilder (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -885,7 +885,7 @@ | Gemini-2.5-NT | 48 | 52,527 | | Qwen3-235b | 50 | 24,125 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -897,9 +897,9 @@ | Gemini-2.5-NT | 2 | 223,682 | | Qwen3-235b | 0 | - | -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) +#### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 22,665 | - | - | | GPT-4o-mini | 32,909 | - | - | @@ -909,47 +909,9 @@ | Gemini-2.5-NT | 52,527 | 223,682 | +171,154 (325.8%) | | Qwen3-235b | 24,125 | - | - | -### intelligent_recruitment_platform (aggregated across architectures) - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 180 | 32,358 | -| GPT-4o-mini | 180 | 92,274 | -| DeepSeek-V3.1 | 180 | 65,486 | -| DeepSeek-R1 | 171 | 77,784 | -| Gemini-2.5 | 180 | 63,570 | -| Gemini-2.5-NT | 180 | 48,983 | -| Qwen3-235b | 133 | 69,503 | - -### Pass (with retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 0 | - | -| DeepSeek-R1 | 8 | 108,468 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 0 | - | -| Qwen3-235b | 44 | 132,115 | - -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) - -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 32,358 | - | - | -| GPT-4o-mini | 92,274 | - | - | -| DeepSeek-V3.1 | 65,486 | - | - | -| DeepSeek-R1 | 77,784 | 108,468 | +30,683 (39.4%) | -| Gemini-2.5 | 63,570 | - | - | -| Gemini-2.5-NT | 48,983 | - | - | -| Qwen3-235b | 69,503 | 132,115 | +62,612 (90.1%) | - -### landing_page_generator (aggregated across architectures) +### LandingPageGenerator (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -961,7 +923,7 @@ | Gemini-2.5-NT | 2 | 188,992 | | Qwen3-235b | 109 | 17,466 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -973,9 +935,9 @@ | Gemini-2.5-NT | 0 | - | | Qwen3-235b | 5 | 33,835 | -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) +#### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 17,805 | - | - | | GPT-4o-mini | 27,799 | - | - | @@ -985,9 +947,9 @@ | Gemini-2.5-NT | 188,992 | - | - | | Qwen3-235b | 17,466 | 33,835 | +16,369 (93.7%) | -### markdown_validator (aggregated across architectures) +### MarkdownValidator (aggregated across variants) -### Pass (no retries) +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -999,7 +961,7 @@ | Gemini-2.5-NT | 113 | 27,744 | | Qwen3-235b | 120 | 1,924 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -1011,9 +973,9 @@ | Gemini-2.5-NT | 7 | 351,232 | | Qwen3-235b | 0 | - | -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) +#### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 1,910 | - | - | | GPT-4o-mini | 1,897 | - | - | @@ -1023,9 +985,85 @@ | Gemini-2.5-NT | 27,744 | 351,232 | +323,488 (1166.0%) | | Qwen3-235b | 1,924 | - | - | -### self_evaluation_loop_flow (aggregated across architectures) +### RecruitmentAssistant (aggregated across variants) + +### Direct Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 180 | 32,358 | +| GPT-4o-mini | 180 | 92,274 | +| DeepSeek-V3.1 | 180 | 65,486 | +| DeepSeek-R1 | 171 | 77,784 | +| Gemini-2.5 | 180 | 63,570 | +| Gemini-2.5-NT | 180 | 48,983 | +| Qwen3-235b | 133 | 69,503 | + +### Retry Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 0 | - | +| GPT-4o-mini | 0 | - | +| DeepSeek-V3.1 | 0 | - | +| DeepSeek-R1 | 8 | 108,468 | +| Gemini-2.5 | 0 | - | +| Gemini-2.5-NT | 0 | - | +| Qwen3-235b | 44 | 132,115 | + +#### Direct Success vs Retry Success (mean, abs & %) + +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 32,358 | - | - | +| GPT-4o-mini | 92,274 | - | - | +| DeepSeek-V3.1 | 65,486 | - | - | +| DeepSeek-R1 | 77,784 | 108,468 | +30,683 (39.4%) | +| Gemini-2.5 | 63,570 | - | - | +| Gemini-2.5-NT | 48,983 | - | - | +| Qwen3-235b | 69,503 | 132,115 | +62,612 (90.1%) | + +### SQLAssistant (aggregated across variants) + +### Direct Success -### Pass (no retries) +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 101 | 7,480 | +| GPT-4o-mini | 114 | 13,910 | +| DeepSeek-V3.1 | 70 | 13,760 | +| DeepSeek-R1 | 36 | 33,567 | +| Gemini-2.5 | 33 | 15,045 | +| Gemini-2.5-NT | 45 | 11,328 | +| Qwen3-235b | 112 | 12,176 | + +### Retry Success + +| Model | n | Mean | +| --- | --- | --- | +| GPT-5 | 79 | 16,279 | +| GPT-4o-mini | 66 | 27,437 | +| DeepSeek-V3.1 | 110 | 34,291 | +| DeepSeek-R1 | 144 | 94,256 | +| Gemini-2.5 | 57 | 37,607 | +| Gemini-2.5-NT | 30 | 39,861 | +| Qwen3-235b | 66 | 23,950 | + +#### Direct Success vs Retry Success (mean, abs & %) + +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | +| --- | --- | --- | --- | +| GPT-5 | 7,480 | 16,279 | +8,799 (117.6%) | +| GPT-4o-mini | 13,910 | 27,437 | +13,527 (97.2%) | +| DeepSeek-V3.1 | 13,760 | 34,291 | +20,532 (149.2%) | +| DeepSeek-R1 | 33,567 | 94,256 | +60,689 (180.8%) | +| Gemini-2.5 | 15,045 | 37,607 | +22,563 (150.0%) | +| Gemini-2.5-NT | 11,328 | 39,861 | +28,532 (251.9%) | +| Qwen3-235b | 12,176 | 23,950 | +11,774 (96.7%) | + +### SocialMediaManager (aggregated across variants) + +### Direct Success | Model | n | Mean | | --- | --- | --- | @@ -1037,7 +1075,7 @@ | Gemini-2.5-NT | 153 | 9,820 | | Qwen3-235b | 97 | 23,270 | -### Pass (with retries) +### Retry Success | Model | n | Mean | | --- | --- | --- | @@ -1049,9 +1087,9 @@ | Gemini-2.5-NT | 27 | 16,691 | | Qwen3-235b | 80 | 65,900 | -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) +#### Direct Success vs Retry Success (mean, abs & %) -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | +| Model | Direct Success mean | Retry Success mean | Δ vs Direct Success | | --- | --- | --- | --- | | GPT-5 | 19,760 | 53,963 | +34,202 (173.1%) | | GPT-4o-mini | 36,934 | 127,046 | +90,112 (244.0%) | @@ -1059,42 +1097,4 @@ | DeepSeek-R1 | 42,398 | 145,943 | +103,544 (244.2%) | | Gemini-2.5 | 12,219 | 27,172 | +14,953 (122.4%) | | Gemini-2.5-NT | 9,820 | 16,691 | +6,872 (70.0%) | -| Qwen3-235b | 23,270 | 65,900 | +42,630 (183.2%) | - -### write_a_book_with_flows (aggregated across architectures) - -### Pass (no retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 180 | 163,074 | -| GPT-4o-mini | 180 | 125,190 | -| DeepSeek-V3.1 | 179 | 198,154 | -| DeepSeek-R1 | 163 | 156,596 | -| Gemini-2.5 | 180 | 102,955 | -| Gemini-2.5-NT | 179 | 78,117 | -| Qwen3-235b | 179 | 101,880 | - -### Pass (with retries) - -| Model | n | Mean | -| --- | --- | --- | -| GPT-5 | 0 | - | -| GPT-4o-mini | 0 | - | -| DeepSeek-V3.1 | 1 | 398,139 | -| DeepSeek-R1 | 17 | 288,873 | -| Gemini-2.5 | 0 | - | -| Gemini-2.5-NT | 1 | 94,226 | -| Qwen3-235b | 1 | 101,570 | - -#### Pass (no retries) vs Pass (with retries) (mean, abs & %) - -| Model | Pass (no retries) mean | Pass (with retries) mean | Δ vs Pass (no retries) | -| --- | --- | --- | --- | -| GPT-5 | 163,074 | - | - | -| GPT-4o-mini | 125,190 | - | - | -| DeepSeek-V3.1 | 198,154 | 398,139 | +199,985 (100.9%) | -| DeepSeek-R1 | 156,596 | 288,873 | +132,277 (84.5%) | -| Gemini-2.5 | 102,955 | - | - | -| Gemini-2.5-NT | 78,117 | 94,226 | +16,109 (20.6%) | -| Qwen3-235b | 101,880 | 101,570 | -310 (-0.3%) | \ No newline at end of file +| Qwen3-235b | 23,270 | 65,900 | +42,630 (183.2%) | \ No newline at end of file diff --git a/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_mix_total_tokens_violin.pdf b/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_mix_total_tokens_violin.pdf index 797d9d6cd8d93c1d31a135d2a54c2887dc8e70a0..c4165292768dbbc1ca7d0eebe6026e9ae82fc777 100644 --- a/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_mix_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_mix_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:97c6fe69e7ca032bf8879e99dc4ba7bf3620de05713bc9567d96c2b8a3e85f03 -size 41225 +oid sha256:a0dd692f8db4a33e77f7d929efb0ff1a36cca550e7d12dd758923d2f9f2ad5eb +size 41479 diff --git a/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_total_tokens_violin.pdf b/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_total_tokens_violin.pdf index e1b612fbfb325df719ade8f4a63cd3589b303424..055d870512bfc7636e792ece290ebde5920c8213 100644 --- a/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/self_evaluation_loop_flow-A2A_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fdb6a47e009351c38ded4816b3a287cff9cfba4b3e1578c38d4eabcbf08cd0c9 -size 44921 +oid sha256:12daa50b440fd797def8f9634e252b308f7834f25ac54af0a9270a79db5e107e +size 45169 diff --git a/data/processed/RQ3/Violin/self_evaluation_loop_flow-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/self_evaluation_loop_flow-MCP_total_tokens_violin.pdf index 5ea0b973ce9277d749a807a846924a4ae2eaaa34..70017a45632c33110fc5762053291dc28483453c 100644 --- a/data/processed/RQ3/Violin/self_evaluation_loop_flow-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/self_evaluation_loop_flow-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:022984cb28abc5e74a74caa6e253982385b99c4b527a3bfef7764b5659f89ffe -size 43795 +oid sha256:2687f6480d63d6f4a9ed3a82cd9cf602e02e378b43f9201ef5e1f472f2a2e1ec +size 44043 diff --git a/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_mix_total_tokens_violin.pdf b/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_mix_total_tokens_violin.pdf index 84c22ba6e417194fa77867cb297c397d247e8eec..fcb69f7fc4f0b1492530fdcf341fa0294b181b33 100644 --- a/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_mix_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_mix_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:06b0bdfa2cdaa7fb0a154a5db99e47004c835e5c573fee09b247a5194522e95d -size 28065 +oid sha256:8cf67830d8e85e6bcc0d7a057385758f45c7c9355fa978d7cc1e32d3b84931a1 +size 28331 diff --git a/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_total_tokens_violin.pdf b/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_total_tokens_violin.pdf index 1c7426c572984561b4521a681adeca5f43d8cdc9..c177a637ea587564dd1f01841d702034a6c92e6e 100644 --- a/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/write_a_book_with_flows-A2A_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:cdbf01cb7a5ed3f65e7d0e770d63e87e3a915c1ec272560def072e834c00b8eb -size 29901 +oid sha256:921a84c06e55889f8a73ffc60fb74ba3b23aa10022596a1d1963232adc47c7de +size 30168 diff --git a/data/processed/RQ3/Violin/write_a_book_with_flows-MCP_total_tokens_violin.pdf b/data/processed/RQ3/Violin/write_a_book_with_flows-MCP_total_tokens_violin.pdf index fe36ccb8d1287a2f81af1954e623109feec5b202..ed75d9308231dff8f94fffa0810c1e9b0f26e7d0 100644 --- a/data/processed/RQ3/Violin/write_a_book_with_flows-MCP_total_tokens_violin.pdf +++ b/data/processed/RQ3/Violin/write_a_book_with_flows-MCP_total_tokens_violin.pdf @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d92f2a944f0cec0129adecf88e3be7cba548406d6dc53aadcb6ad8ccaeac09f5 -size 29973 +oid sha256:aa301b135f1f46b17e27dde1ebcc0c2d0fde9fef545d3eca3612217c0b6057d5 +size 30238 diff --git a/data/processed/RQ3/analyze_performance.py b/data/processed/RQ3/analyze_performance.py index 0660d89241654a2e9e7ca1cc3b74bc1d4e7eb9f8..12fc032d66fd6974463d492fbc9770153cec41d8 100644 --- a/data/processed/RQ3/analyze_performance.py +++ b/data/processed/RQ3/analyze_performance.py @@ -25,6 +25,24 @@ logging.basicConfig( logger = logging.getLogger(__name__) +def normalize_task_name(task: Optional[str]) -> str: + raw = (task or "").strip() + if not raw: + return "" + + # Map the actual project/task directory names to the canonical task names + # used by success checking logic. + mapping = { + "SocialMediaManager": "self_evaluation_loop_flow", + "SQLAssistant": "SQL_assistant", + "LandingPageGenerator": "landing_page_generator", + "RecruitmentAssistant": "intelligent_recruitment_platform", + "BookWriter": "write_a_book_with_flows", + } + + return mapping.get(raw, raw) + + class ExecutionPathParser: """Parser for execution_path.md files""" @@ -70,16 +88,21 @@ class ExecutionPathParser: # Extract architecture from task name if "-MCP" in task_full: self.architecture = "MCP" - self.task = task_full.replace("-MCP", "") + self.task = normalize_task_name(task_full.replace("-MCP", "")) elif "-A2A_mix" in task_full: self.architecture = "A2A_mix" - self.task = task_full.replace("-A2A_mix", "") + self.task = normalize_task_name(task_full.replace("-A2A_mix", "")) + elif "-H_A2A" in task_full or "-H-A2A" in task_full: + self.architecture = "A2A_mix" + self.task = normalize_task_name( + task_full.replace("-H_A2A", "").replace("-H-A2A", "") + ) elif "-A2A" in task_full: self.architecture = "A2A" - self.task = task_full.replace("-A2A", "") + self.task = normalize_task_name(task_full.replace("-A2A", "")) else: # Check in file content for architecture info - self.task = task_full + self.task = normalize_task_name(task_full) self.architecture = "Unknown" logger.debug( @@ -599,9 +622,10 @@ class PerformanceAnalyzer: # Get the directory containing execution_path.md session_dir = Path(parser.file_path).parent task = parser.task + task_norm = normalize_task_name(task) # 1. write_a_book_with_flows - if task == "write_a_book_with_flows": + if task_norm == "write_a_book_with_flows": chapters_dir = session_dir / "chapters" if chapters_dir.exists() and chapters_dir.is_dir(): md_files = list(chapters_dir.glob("*.md")) @@ -610,7 +634,7 @@ class PerformanceAnalyzer: return "fail" # 2. SQL_assistant - elif task == "SQL_assistant": + elif task_norm == "SQL_assistant": with open(parser.file_path, "r", encoding="utf-8") as f: content = f.read() if re.search(r"\[Tool\]\s+get_database_schema", content): @@ -618,7 +642,7 @@ class PerformanceAnalyzer: return "fail" # 3. self_evaluation_loop_flow - elif task == "self_evaluation_loop_flow": + elif task_norm == "self_evaluation_loop_flow": metadata_file = session_dir / "metadata.json" if metadata_file.exists(): with open(metadata_file, "r", encoding="utf-8") as f: @@ -628,7 +652,7 @@ class PerformanceAnalyzer: return "fail" # 4. MarkdownValidator - elif task == "MarkdownValidator": + elif task_norm == "MarkdownValidator": execution_info_file = session_dir / "execution_info.json" if execution_info_file.exists(): with open(execution_info_file, "r", encoding="utf-8") as f: @@ -638,7 +662,7 @@ class PerformanceAnalyzer: return "fail" # 5. landing_page_generator - elif task == "landing_page_generator": + elif task_norm == "landing_page_generator": html_validation_file = session_dir / "html_validation.json" if html_validation_file.exists(): with open(html_validation_file, "r", encoding="utf-8") as f: @@ -648,7 +672,7 @@ class PerformanceAnalyzer: return "fail" # 6. intelligent_recruitment_platform - elif task == "intelligent_recruitment_platform": + elif task_norm == "intelligent_recruitment_platform": reports_dir = session_dir / "reports" if reports_dir.exists() and reports_dir.is_dir(): md_files = list(reports_dir.glob("*.md")) @@ -657,7 +681,7 @@ class PerformanceAnalyzer: return "fail" # 7. GameBuilder - elif task == "GameBuilder": + elif task_norm == "GameBuilder": validation_result_file = session_dir / "validation_result.json" if validation_result_file.exists(): with open(validation_result_file, "r", encoding="utf-8") as f: @@ -667,7 +691,7 @@ class PerformanceAnalyzer: return "fail" # 8. EmailResponder - elif task == "EmailResponder": + elif task_norm == "EmailResponder": execution_log_file = session_dir / "execution_log.json" if execution_log_file.exists(): with open(execution_log_file, "r", encoding="utf-8") as f: @@ -678,7 +702,9 @@ class PerformanceAnalyzer: # Unknown task else: - logger.warning(f"Unknown task type for success check: {task}") + logger.warning( + f"Unknown task type for success check: task={task_norm}, session_dir={session_dir}" + ) return "unknown" except Exception as e: @@ -692,11 +718,10 @@ class PerformanceAnalyzer: results = [] for parser in self.data: - if parser.total_tokens["total"] > 0: - # Check task success status + if parser.total_tokens["total"] <= 0: + continue + try: status = self.check_task_success(parser) - - # Check if execution contains retry markers has_retry = self.check_has_retry(parser) result = { @@ -705,9 +730,7 @@ class PerformanceAnalyzer: "architecture": parser.architecture, "model": parser.model, "status": status, - "with_retry": str( - has_retry - ).lower(), # Convert to 'true' or 'false' + "with_retry": str(has_retry).lower(), "input_tokens": parser.total_tokens["input"], "output_tokens": parser.total_tokens["output"], "reasoning_tokens": parser.total_tokens["reasoning"], @@ -715,12 +738,21 @@ class PerformanceAnalyzer: "total_tokens": parser.total_tokens["total"], } results.append(result) + except Exception as e: + logger.error( + f"Error generating token details for {getattr(parser, 'file_path', '')}: {e}", + exc_info=True, + ) # Sort by task, architecture, model, and file path results.sort( key=lambda x: (x["task"], x["architecture"], x["model"], x["file_path"]) ) + out_dir = os.path.dirname(output_file) + if out_dir: + os.makedirs(out_dir, exist_ok=True) + # Write to CSV with open(output_file, "w", newline="", encoding="utf-8") as f: writer = csv.DictWriter( diff --git a/data/processed/RQ3/plot_total_tokens_violin.py b/data/processed/RQ3/plot_total_tokens_violin.py index 625dfb17b36a42c0ce44e50d7aa0e9177a577973..3eb003f242f1db7d9d4cf4a5e404cffe6e23aad6 100644 --- a/data/processed/RQ3/plot_total_tokens_violin.py +++ b/data/processed/RQ3/plot_total_tokens_violin.py @@ -66,8 +66,8 @@ STATUS_ORDER: List[str] = [ ] STATUS_TITLES: Dict[str, str] = { - "success_no_retry": "Pass (no retries)", - "success_with_retry": "Pass (with retries)", + "success_no_retry": "Direct Success", + "success_with_retry": "Retry Success", "failed": "Failure", } @@ -272,13 +272,57 @@ def load_projects( def classify_status(status_raw: str, with_retry_raw: str) -> str: s = (status_raw or "").strip().lower() w = (with_retry_raw or "").strip().lower() - if s == "success" and w == "false": + if s in {"success", "pass", "passed", "ok"} and w == "false": return "success_no_retry" - if s == "success" and w == "true": + if s in {"success", "pass", "passed", "ok"} and w == "true": return "success_with_retry" return "failed" +def select_details_csv(part1_dir: Path) -> Path: + candidates = [ + part1_dir / "task_token_statistics-DETAILS.csv", + part1_dir / "performance_reports" / "task_token_statistics-DETAILS.csv", + Path( + "/Users/wzr/TOSEM-2025/RESULTS/RQ3/performance_reports/task_token_statistics-DETAILS.csv" + ), + ] + + existing = [p for p in candidates if p.exists()] + if not existing: + return candidates[0] + + def score_csv(path: Path, sample_rows: int = 500) -> Tuple[float, float, float]: + """Return (success_ratio, -unknown_ratio, mtime) for tie-breaking.""" + total = 0 + success = 0 + unknown = 0 + try: + with path.open("r", encoding="utf-8", newline="") as f: + reader = csv.DictReader(f) + for row in reader: + if total >= sample_rows: + break + total += 1 + s = (row.get("status") or "").strip().lower() + if s == "success": + success += 1 + if s == "unknown": + unknown += 1 + except Exception: + total = 0 + success = 0 + unknown = 0 + + if total <= 0: + return (0.0, 0.0, float(path.stat().st_mtime)) + success_ratio = success / float(total) + unknown_ratio = unknown / float(total) + return (success_ratio, -unknown_ratio, float(path.stat().st_mtime)) + + return max(existing, key=score_csv) + + def load_total_tokens( details_csv: Path, project_map: Dict[Tuple[str, str], str] ) -> Dict[str, Dict[str, Dict[str, List[float]]]]: @@ -979,12 +1023,15 @@ def plot_violin_for_project( def main() -> None: part1_dir = Path(__file__).resolve().parent - details_csv = part1_dir / "task_token_statistics-DETAILS.csv" + details_csv = select_details_csv(part1_dir) + out_dir = part1_dir / "Violin" + if not details_csv.exists(): - details_csv = ( - part1_dir / "performance_reports" / "task_token_statistics-DETAILS.csv" + raise FileNotFoundError( + f"task_token_statistics-DETAILS.csv not found: {details_csv}" ) - out_dir = part1_dir / "Violin" + + print(f"using token details: {details_csv}") projects, project_map = load_projects(details_csv)