Loom / benchmarks.json
deploy-space action
deploy 6158a00 (c)
b972454
Raw
History Blame Contribute Delete
3.27 kB
{
"//": "Indicative, HAND-CURATED quality signal per logical model (as of 2026-06).",
"//1": "NOT an authoritative ranking - arena_elo/aa_index are approximate, sourced from",
"//2": "public leaderboards (LMArena, Artificial Analysis) and rounded. Used only to",
"//3": "ORDER the roster and decide which models count toward the >=2-frontier guarantee.",
"//4": "'frontier': true marks a top-tier model for that guarantee. Update with sources.",
"benchmarks": {
"qwen3.5-397b": {"arena_elo": 1378, "aa_index": 67, "frontier": true, "reliability": "flaky-free-tier (returned empty content on a 163k-token review 2026-06; demoted from default panel)", "source": "qwen3.5 report MMLU-Pro 84.7 (approx)", "as_of": "2026-06"},
"deepseek-v4-flash": {"arena_elo": 1372, "aa_index": 65, "frontier": true, "source": "v4 distill, probe-verified (approx)", "as_of": "2026-06"},
"deepseek-v4-pro": {"arena_elo": 1380, "aa_index": 68, "frontier": false, "source": "strong but DEMOTED (queue >1k)", "as_of": "2026-06"},
"qwen3.5-122b": {"arena_elo": 1342, "aa_index": 60, "frontier": true, "source": "qwen3.5 report MMLU-Pro 80.4 (approx)", "as_of": "2026-06"},
"mistral-medium-3.5":{"arena_elo": 1332, "aa_index": 59, "frontier": true, "source": "mistral blog MMLU-Pro 79.1 (approx)", "as_of": "2026-06"},
"llama-4-maverick": {"arena_elo": 1326, "aa_index": 58, "frontier": true, "source": "llama 4 blog MMLU-Pro 78.5 (approx)", "as_of": "2026-06"},
"minimax-m2.7": {"arena_elo": 1322, "aa_index": 57, "frontier": true, "source": "minimax card MMLU-Pro 77.8 (approx)", "as_of": "2026-06"},
"mistral-small-4": {"arena_elo": 1288, "aa_index": 53, "frontier": false, "source": "mistral card MMLU-Pro 74.8 (approx)", "as_of": "2026-06"},
"kimi-k2.6": {"arena_elo": 1365, "aa_index": 66, "frontier": true, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"glm-5.1": {"arena_elo": 1355, "aa_index": 64, "frontier": true, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"nemotron-ultra": {"arena_elo": 1345, "aa_index": 63, "frontier": true, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"deepseek-v3": {"arena_elo": 1335, "aa_index": 60, "frontier": true, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"deepseek-r1": {"arena_elo": 1330, "aa_index": 60, "frontier": true, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"gpt-oss-120b": {"arena_elo": 1320, "aa_index": 58, "frontier": true, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"nemotron-120b": {"arena_elo": 1300, "aa_index": 55, "frontier": true, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"step-flash": {"arena_elo": 1280, "aa_index": 52, "frontier": false, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"phi-4-reasoning": {"arena_elo": 1265, "aa_index": 50, "frontier": false, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"llama-3.3-70b": {"arena_elo": 1255, "aa_index": 48, "frontier": false, "source": "lmarena+aa (approx)", "as_of": "2026-06"},
"gemma-4": {"arena_elo": 1230, "aa_index": 45, "frontier": false, "source": "lmarena+aa (approx)", "as_of": "2026-06"}
}
}