Datasets:

psinglap
/

websearch-api-eval

 ---
 license: mit
+pretty_name: Web Search API Eval
+language:
+- en
+size_categories:
+- n<1K
+task_categories:
+- question-answering
+- text-retrieval
+tags:
+- web-search
+- search-api
+- benchmark
+- evaluation
+- rag
+- ai-agents
+- freshness
+- you.com
+- tavily
+- exa
+- parallel
+configs:
+- config_name: questions
+  data_files: questions.jsonl
+  default: true
+- config_name: results
+  data_files: results.jsonl
+- config_name: scorecards
+  data_files: scorecards.jsonl
 ---
+# Web Search API Eval
+A live, side-by-side benchmark of four web search APIs for AI agents: **You.com, Tavily, Exa and Parallel**. It covers 30 agent-style queries across three use cases (GTM, insurance, finance), with 3 runs per query (90 calls per API), run on **October 8, 2026**.
+- **Full results and charts:** https://psinglap.github.io/websearch-api-eval/
+- **Code, harness and methodology:** https://github.com/psinglap/websearch-api-eval
+- **Use it from Claude, ChatGPT or Cursor (MCP):** `https://websearch-api-eval.psberkeleymba.workers.dev/mcp` or `npx -y websearch-api-eval-mcp`
+## Subsets
+| Config | Rows | What each row is |
+|---|---|---|
+| `questions` (default) | 30 | One agent query: the upstream task, the exact search call sent to every API, its parameters (freshness window, domain filters), what the agent needs back, the answer key captured live, and source URLs |
+| `results` | 120 | One query × provider: a short summary of what the API returned and its accuracy score (0–2, mean of 3 runs, SimpleQA rubric) |
+| `scorecards` | 136 | One metric × provider × scope (`overall`, `gtm`, `insurance`, `finance`): latency, accuracy, freshness, empty responses, content, tokens and cost, as published |
+```python
+from datasets import load_dataset
+questions = load_dataset("psinglap/websearch-api-eval", "questions", split="train")
+results = load_dataset("psinglap/websearch-api-eval", "results", split="train")
+```
+## Headline results (overall, 90 calls per API)
+| Metric | You.com | Tavily | Exa | Parallel |
+|---|---|---|---|---|
+| Accuracy score (% of max) | 63% | 51% | 53% | **67%** |
+| Queries fully correct in all 3 repeats | **57%** | 37% | 50% | **57%** |
+| Results dated inside the query's window | **59%** | 37% | 44% | 29% |
+| Empty responses (of 90) | **0** | 27 | 12 | 3 |
+| Latency p50 / p95, client (ms) | **593 / 804** | 1,864 / 3,305 | 1,106 / 3,997 | 2,038 / 4,117 |
+| Search cost per correct answer | $0.026 | $0.016 | $0.012 | **$0.008** |
+## Methodology
+- **Queries.** Each query is written the way an agent calls web search: an upstream task, the search call, its parameters and what the agent needs back. Every use case has at least 4 real-time queries; the rest use 24-hour, 48-hour, 7-day or 1-year windows. Some queries expect "none" as the correct answer.
+- **Same request for every API.** 10 results, full content where supported, the freshness window, and domain filters where supported.
+- **Runs.** promptfoo, 3 runs per query. Latency is client wall-clock around the HTTP request after a discarded warm-up call; server-reported latency is kept alongside.
+- **Grading.** Answers were graded from each provider's saved top-5 results with the published [SimpleQA](https://arxiv.org/abs/2411.04368) grader rubric (correct = 2, partial list = 1, incorrect or not attempted = 0). "None" answers and stale results follow [FreshQA](https://arxiv.org/abs/2310.03214) strict rules. Answer keys were captured at run time and archived to the Wayback Machine.
+- **Full metric definitions:** [eval/METRICS.md](https://github.com/psinglap/websearch-api-eval/blob/main/eval/METRICS.md).
+## Limits
+- 10 queries per use case is a focused test, not a leaderboard; differences of a query or two are within noise.
+- Results reflect one run day (Oct 8, 2026). Freshness and accuracy depend on what was published that day, so answer keys for real-time queries are only valid for that date.
+- Costs are vendor list prices on the run date and cover search only.
+- `returned` is a short human-written summary, not the raw API response.
+## Citation
+```bibtex
+@misc{singla2026websearchapieval,
+  title  = {Web Search API Eval: You.com, Tavily, Exa and Parallel on 30 agent queries},
+  author = {Singla, Pallavi},
+  year   = {2026},
+  url    = {https://github.com/psinglap/websearch-api-eval}
+}
+```
+## License
+MIT. Product names and trademarks belong to their owners. This benchmark is independent and was not sponsored by any provider.

questions.jsonl ADDED Viewed

	@@ -0,0 +1,30 @@

+{"id": "G1", "use_case": "gtm", "task": "Find developers hitting Kafka latency problems right now", "search_call": "kafka consumer lag high latency issue", "parameters": "window last ~6 h (tightest available), domains: github.com", "needs_back": "Issue URLs opened in the last 6 hours, repo/org, time, the error described", "freshness": "Real-time", "answer_key": "Brighter#4560 (opened 10:51Z, Kafka publish confirms delay under load) or NOFireAI/opentelemetry-demo#7024 (15:50Z, Kafka consumer-lag spike). Issues opened Oct 7 are out of window.", "source_urls": ["https://github.com/BrighterCommand/Brighter/issues/4560", "https://github.com/search?q=kafka+consumer+lag+latency&type=issues&s=created&o=desc", "https://web.archive.org/web/20261008154559/https://github.com/search?q=kafka+consumer+lag+latency&type=issues&s=created&o=desc"]}
+{"id": "G2", "use_case": "gtm", "task": "Account signal: anything new at Confluent today?", "search_call": "Confluent news", "parameters": "window 24h", "needs_back": "Headline, date, link or \"none\" in the last 24h", "freshness": "Real-time", "answer_key": "All Things Open interview with Confluent VP Tim Berglund, dated Oct 8. No new official Confluent blog post in 24h (newest Oct 1); undated pages and the IBM deal do not count. \"None\" is wrong.", "source_urls": ["https://allthingsopen.org/articles/ai-strategy-data-strategy-confluent", "https://www.confluent.io/blog/", "https://web.archive.org/web/20261008145909/https://www.confluent.io/blog/"]}
+{"id": "G3", "use_case": "gtm", "task": "Any new mentions of Apache Iggy in the last day?", "search_call": "\"Apache Iggy\"", "parameters": "window 24h", "needs_back": "Every mention (repo, blog, forum) in 24h, date or \"none\"", "freshness": "Real-time", "answer_key": "32 new apache/iggy issues and PRs in window, e.g. #4466 (13:37Z), #4443 (10:53Z), #4438 (08:31Z). No blog or HN mention. \"None\" is wrong.", "source_urls": ["https://github.com/apache/iggy", "https://web.archive.org/web/20261008150545/https://github.com/apache/iggy"]}
+{"id": "G4", "use_case": "gtm", "task": "Draft a note on Apache Iggy's latest release for a prospect", "search_call": "Apache Iggy release notes", "parameters": "window 7 days, domains: github.com, iggy.apache.org", "needs_back": "Latest release version, date, key changes, link", "freshness": "Week", "answer_key": "server-0.9.0, released Sep 18 (GitHub \"Latest\"): consensus per partition (VSR clustering), QUIC/TLS/WS transports, Delta Lake sink, Avro. The downloads page's 0.5.0-incubating is wrong.", "source_urls": ["https://github.com/apache/iggy/releases/tag/server-0.9.0", "https://github.com/apache/iggy/releases", "https://web.archive.org/web/20261008150847/https://github.com/apache/iggy/releases"]}
+{"id": "G5", "use_case": "gtm", "task": "Coding agent fixing a consumer-lag bug needs the official tuning guidance", "search_call": "kafka consumer lag max.poll.records fetch.min.bytes", "parameters": "window 1 year, domains: kafka.apache.org, docs.confluent.io", "needs_back": "Official doc URL, what each setting does, recommended direction", "freshness": "Year", "answer_key": "Official Kafka or Confluent consumer-config page. max.poll.records = max records per poll() (default 500); lower it if processing risks max.poll.interval.ms. fetch.min.bytes = min data per fetch (default 1 byte); raising it trades latency for throughput. Both settings needed.", "source_urls": ["https://docs.confluent.io/platform/current/installation/configuration/consumer-configs.html", "https://web.archive.org/web/20261008151030/https://docs.confluent.io/platform/current/installation/configuration/consumer-configs.html"]}
+{"id": "G6", "use_case": "gtm", "task": "Competitive brief: what's the latest NATS server release?", "search_call": "NATS server release", "parameters": "window 7 days, domains: github.com", "needs_back": "Version, release date, link", "freshness": "Week", "answer_key": "v2.15.1-RC.2 (pre-release, Oct 6) or v2.15.0 (latest stable, Sep 17) both count; an older version as latest is wrong.", "source_urls": ["https://github.com/nats-io/nats-server/releases/tag/v2.15.1-RC.2", "https://github.com/nats-io/nats-server/releases/tag/v2.15.0", "https://github.com/nats-io/nats-server/releases", "https://web.archive.org/web/20261008111533/https://github.com/nats-io/nats-server/releases"]}
+{"id": "G7", "use_case": "gtm", "task": "Is Confluent Cloud having an incident right now? (outage = switching trigger)", "search_call": "Confluent Cloud incident", "parameters": "window 24h, domains: status.confluent.cloud", "needs_back": "Active incident yes/no, component, start time or \"none\"", "freshness": "Real-time", "answer_key": "None: no active incident (unresolved.json empty, All Systems Operational at 15:51-15:59Z). The Azure Germany West Central incident (11:00-11:56Z) was resolved; showing it or the Sep 21 incident as current is wrong.", "source_urls": ["https://status.confluent.cloud/", "https://web.archive.org/web/20261008155137/https://status.confluent.cloud/api/v2/incidents/unresolved.json"]}
+{"id": "G8", "use_case": "gtm", "task": "Which AI-agent / data-infra startups just raised money?", "search_call": "AI agent infrastructure startup funding round", "parameters": "window 48h", "needs_back": "Company, amount, lead investor, date, link", "freshness": "48h", "answer_key": "Open-world: any AI-agent or data-infra round announced in the 48h window with company + amount, e.g. Vesta $30M and Mecka AI $60M (TechCrunch, Oct 8). Rounds before Oct 6 15:52Z (Namespace $42M) are out of window.", "source_urls": ["https://techcrunch.com/category/venture/", "https://web.archive.org/web/20261008111803/https://techcrunch.com/category/venture/"]}
+{"id": "G9", "use_case": "gtm", "task": "Find this week's content about memory/event logs for AI agents", "search_call": "durable event log memory for AI agents", "parameters": "window 7 days", "needs_back": "Article URLs, author/org, date", "freshness": "Week", "answer_key": "Open-world: 126 HN items in window, e.g. \"Agents Don't Need Another Memory API\" (Oct 7); any dated in-window article counts.", "source_urls": ["https://hn.algolia.com/?query=agent%20memory&sort=byDate", "https://web.archive.org/web/20261008111956/https://hn.algolia.com/?query=agent%20memory&sort=byDate"]}
+{"id": "G10", "use_case": "gtm", "task": "Competitive brief: published Redpanda vs Kafka latency benchmarks", "search_call": "Redpanda vs Kafka p99 latency benchmark", "parameters": "window 1 year", "needs_back": "Benchmark source, date, p99 numbers, who ran it", "freshness": "Year", "answer_key": "Open-world criterion: a benchmark dated in the last year with Redpanda vs Kafka p99 numbers and who ran it. Redpanda's own 2022 benchmark is out of window; no official post in the year.", "source_urls": ["https://www.redpanda.com/blog", "https://web.archive.org/web/20261008151435/https://www.redpanda.com/blog"]}
+{"id": "I1", "use_case": "insurance", "task": "Caller wants a new Florida homeowners policy", "search_call": "Citizens Property Insurance binding suspension Florida", "parameters": "window 24h, domains: citizensfla.com", "needs_back": "Suspended yes/no, scope, effective date/time", "freshness": "Real-time", "answer_key": "Yes: statewide binding suspension (\"Isaias: No Policy Binding\"), no new policies or added coverage anywhere in Florida, effective Oct 7, 2026, 11:05 a.m.; still posted before and after the run.", "source_urls": ["https://www.citizensfla.com/", "https://web.archive.org/web/20261008112205/https://www.citizensfla.com/"]}
+{"id": "I2", "use_case": "insurance", "task": "Caller in Pensacola asks if the storm is coming", "search_call": "Hurricane Isaias latest advisory NHC", "parameters": "window 24h", "needs_back": "Advisory #, time, location, max winds, landfall forecast", "freshness": "Real-time", "answer_key": "Yes, heading for Pensacola: NHC Public Advisory #8, 10:00 AM CDT (15:00Z) Oct 8: 23.7N 90.2W, ~500 mi SSW of Pensacola, 85 mph, 975 mb, moving ENE at 10 mph; landfall on the north-central Gulf coast Friday night. Unchanged before and after the run.", "source_urls": ["https://www.nhc.noaa.gov/", "https://web.archive.org/web/20261008090330/https://www.nhc.noaa.gov/"]}
+{"id": "I3", "use_case": "insurance", "task": "Agent must check warnings for the policy's county before binding", "search_call": "hurricane warning Escambia County Florida active", "parameters": "window 24h", "needs_back": "Active warning types, issued time", "freshness": "Real-time", "answer_key": "Hurricane Warning (Escambia Inland and Coastal) and Storm Surge Warning (Escambia Coastal), NWS Mobile, latest update 10:06 AM CDT Oct 8 (an earlier Oct 7-8 issue time also counts).", "source_urls": ["https://api.weather.gov/alerts/active?zone=FLC033,FLZ201,FLZ202", "https://web.archive.org/web/20261008112313/https://api.weather.gov/alerts/active?zone=FLC033,FLZ201,FLZ202"]}
+{"id": "I4", "use_case": "insurance", "task": "Claims agent validating a wind-damage report in progress", "search_call": "Pensacola weather wind speed", "parameters": "window 24h", "needs_back": "Current wind speed/gusts, observation time", "freshness": "Real-time", "answer_key": "Pensacola Intl (KPNS): 15:50Z observation 10.4 mph NE, gusts 16.1 mph; 16:00Z observation 9.2 mph ENE, no gust. Either counts (within 5 mph, at most 1 h old); the hurricane was still ~500 mi offshore.", "source_urls": ["https://api.weather.gov/stations/KPNS/observations/latest"]}
+{"id": "I5", "use_case": "insurance", "task": "Policyholder asks if they must evacuate", "search_call": "Baldwin County Alabama evacuation order Isaias", "parameters": "window 24h", "needs_back": "Order in effect yes/no, zones, time", "freshness": "Real-time", "answer_key": "None: no mandatory evacuation order (a voluntary evacuation for coastal and low-lying areas is compatible); Hurricane and Storm Surge Warnings in effect. Moderate confidence: EMA sub-pages not captured.", "source_urls": ["https://baldwincountyal.gov/departments/EMA", "https://api.weather.gov/alerts/active?zone=ALC003", "https://web.archive.org/web/20261008112609/https://baldwincountyal.gov/services/ema", "https://web.archive.org/web/20261008112650/https://api.weather.gov/alerts/active?zone=ALC003"]}
+{"id": "I6", "use_case": "insurance", "task": "Claims ops checks whether federal aid applies", "search_call": "FEMA disaster declaration Hurricane Isaias 2026", "parameters": "window 48h", "needs_back": "Declaration number + date — or \"none yet\" (beware older \"Isaias\")", "freshness": "48h", "answer_key": "None yet: all 183 OpenFEMA records titled Isaias are from 2020-21; none in the 48 h window.", "source_urls": ["https://www.fema.gov/api/open/v2/DisasterDeclarationsSummaries", "https://www.fema.gov/disaster/declarations"]}
+{"id": "I7", "use_case": "insurance", "task": "Compliance agent checks for new regulator orders", "search_call": "Florida OIR OR Alabama DOI emergency order Isaias insurance", "parameters": "window 48h", "needs_back": "Order/bulletin, issuer, date — or \"none\"", "freshness": "48h", "answer_key": "None: no Isaias emergency order or bulletin from Florida OIR or Alabama DOI in 48 h; an ALDOI consumer release and an FLOIR refill notice (both Oct 7) are not orders. Moderate confidence: order sub-pages not captured.", "source_urls": ["https://floir.com/", "https://www.aldoi.gov/", "https://web.archive.org/web/20261008113340/https://floir.gov/", "https://web.archive.org/web/20261008113421/https://aldoi.gov/"]}
+{"id": "I8", "use_case": "insurance", "task": "Underwriting checks which states declared emergencies", "search_call": "governor state of emergency Hurricane Isaias", "parameters": "window 7 days", "needs_back": "States, date of each declaration", "freshness": "Week", "answer_key": "Florida: EO 26-202, Oct 6 (25 counties); Alabama: Oct 7. Mississippi: none as of 15:27Z; Louisiana: not determinable.", "source_urls": ["https://www.flgov.com/eog/news/press", "https://governor.alabama.gov/newsroom/", "https://gov.louisiana.gov/news", "https://governor.ms.gov/", "https://web.archive.org/web/20261008113558/https://www.flgov.com/eog/news/press", "https://web.archive.org/web/20261008113637/https://governor.alabama.gov/newsroom/", "https://web.archive.org/web/20261008113716/https://gov.louisiana.gov/news", "https://web.archive.org/web/20261008113757/https://governorreeves.ms.gov/"]}
+{"id": "I9", "use_case": "insurance", "task": "Claims agent validating a Texas hail claim", "search_call": "Texas hail storm reports", "parameters": "window 24h, domains: spc.noaa.gov", "needs_back": "Locations, dates, hail size — or \"none\"", "freshness": "24h", "answer_key": "None: no Texas hail reports in SPC preliminary storm reports for the last 24 h; the only hail report was in Wilson, Niagara Co., NY (1.00 in, 01:30Z).", "source_urls": ["https://www.spc.noaa.gov/climo/reports/today.html", "https://www.spc.noaa.gov/climo/reports/yesterday.html"]}
+{"id": "I10", "use_case": "insurance", "task": "Caller asks whether their 2024 Toyota RAV4 has an open recall", "search_call": "NHTSA recall 2024 Toyota RAV4", "parameters": "window 1 year, domains: nhtsa.gov", "needs_back": "Recall campaign number(s), component, date", "freshness": "Year", "answer_key": "Three NHTSA campaigns: 24V911000 (Dec 2024, front brake caliper bolts), 25V595000 (Sep 2025, instrument cluster software) and 25V744000 (Oct 2025, rearview camera). All three = correct; a subset = partial (26V179000, RAV4 Hybrid load label, also counts).", "source_urls": ["https://api.nhtsa.gov/recalls/recallsByVehicle?make=toyota&model=rav4&modelYear=2024", "https://www.nhtsa.gov/recalls"]}
+{"id": "F1", "use_case": "finance", "task": "Where is NVIDIA trading?", "search_call": "NVDA stock price", "parameters": "window 24h", "needs_back": "Price, % change, timestamp", "freshness": "Real-time", "answer_key": "~$235.6, down ~0.8% on the day: Yahoo $235.58 at 12:24 PM EDT before the run and $235.60 at 12:35 PM after (market open). Within ±0.5% counts; the Oct 7 close of $237.47 is stale.", "source_urls": ["https://finance.yahoo.com/quote/NVDA/", "https://www.nasdaq.com/market-activity/stocks/nvda"]}
+{"id": "F2", "use_case": "finance", "task": "Where's the 10-year?", "search_call": "US 10-year Treasury yield", "parameters": "window 24h", "needs_back": "Yield %, timestamp", "freshness": "Real-time", "answer_key": "~5.27%: CNBC US10Y 5.265% at 12:25 PM EDT and 5.269% at 12:36 PM (prior close 5.277%; a 24-year high of 5.35% earlier in the session). Within ±3 bp counts; morning quotes of 5.31-5.35% do not.", "source_urls": ["https://www.cnbc.com/quotes/US10Y", "https://home.treasury.gov/resource-center/data-chart-center/interest-rates"]}
+{"id": "F3", "use_case": "finance", "task": "Bitcoin price?", "search_call": "BTC USD price", "parameters": "window 24h", "needs_back": "Price, timestamp", "freshness": "Real-time", "answer_key": "~$81,200-81,400: Coinbase spot $81,169 at run start (16:33Z) and $81,376 after (16:41Z), down ~2.5-3% on the day. Within ±1% counts.", "source_urls": ["https://www.coinbase.com/price/bitcoin", "https://api.coinbase.com/v2/prices/BTC-USD/spot"]}
+{"id": "F4", "use_case": "finance", "task": "How's oil moving today?", "search_call": "WTI crude oil price", "parameters": "window 24h", "needs_back": "Price, change, timestamp", "freshness": "Real-time", "answer_key": "~$91.00/bbl, up ~3% on the $88.28 prior close: CNBC @CL.1 $90.95 at 12:20 PM EDT and $91.04 at 12:31 PM (down from ~$92.8 at 11:26 AM). Within ±1% counts.", "source_urls": ["https://www.cnbc.com/quotes/@CL.1"]}
+{"id": "F5", "use_case": "finance", "task": "What did jobless claims print?", "search_call": "initial jobless claims weekly report", "parameters": "window 24h, domains: dol.gov", "needs_back": "Number, week ending, release time", "freshness": "Real-time", "answer_key": "197,000 (seasonally adjusted, advance) for the week ending Oct 3, released Oct 8 at 8:30 a.m. ET; down 2,000 from the revised 199,000. The prior week's print was also 197,000, so the week must be right.", "source_urls": ["https://www.dol.gov/ui/data.pdf", "https://www.dol.gov/newsroom/releases"]}
+{"id": "F6", "use_case": "finance", "task": "Summarize the latest Fed minutes", "search_call": "FOMC minutes", "parameters": "window 7 days, domains: federalreserve.gov", "needs_back": "Release date, 3 key points, federalreserve.gov link", "freshness": "Week", "answer_key": "Minutes of the Sep 15-16, 2026 meeting, released Oct 7 at 2:00 p.m. EDT (fomcminutes20260916). The July minutes (released Aug 19) are stale.", "source_urls": ["https://www.federalreserve.gov/monetarypolicy/fomcminutes20260916.htm", "https://www.federalreserve.gov/monetarypolicy/fomccalendars.htm", "https://web.archive.org/web/20261008095612/https://www.federalreserve.gov/monetarypolicy/fomccalendars.htm"]}
+{"id": "F7", "use_case": "finance", "task": "List $1B+ M&A deals announced in the last two days", "search_call": "definitive agreement to acquire billion", "parameters": "window 48h", "needs_back": "Acquirer, target, value, date, source", "freshness": "48h", "answer_key": "Open list: Viatris-Pacira BioSciences (~$1.65B, Oct 8) and Wittington/Fairfax-Boots (~US$8.9B, Oct 7). Oct 6 deals (RXO, Option Care, Sunrise) are borderline; Schneider-PTC (Oct 5) is out of window.", "source_urls": ["https://efts.sec.gov/LATEST/search-index?q=%22definitive%20agreement%22&forms=8-K"]}
+{"id": "F8", "use_case": "finance", "task": "Monitoring account Stryker: what did its latest 8-K report?", "search_call": "Stryker 8-K item 5.02", "parameters": "window 7 days, domains: sec.gov", "needs_back": "Filing date, items, what changed (officer/director), EDGAR link", "freshness": "Week", "answer_key": "8-K filed Oct 6, 2026 (Items 5.02, 7.01, 9.01): Kevin Lobo steps down as CEO Dec 31 and becomes Executive Chair; Spencer Stiles becomes CEO Jan 1, 2027 (filing index). The May 20 8-K is stale.", "source_urls": ["https://www.sec.gov/Archives/edgar/data/310764/000119312526414956/0001193125-26-414956-index.htm", "https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000310764&type=8-K"]}
+{"id": "F9", "use_case": "finance", "task": "Analyst asks about the Chilwa Minerals listing", "search_call": "Chilwa Minerals IPO price per share", "parameters": "window 7 days", "needs_back": "IPO price, ticker, exchange, pricing date", "freshness": "Week", "answer_key": "US$5.60 per ADS (625,000 ADSs, ~US$3.5M), ticker CHWM on the Nasdaq Capital Market, priced Oct 1, 2026 (release dated Sep 30). The expected $5.00 terms are stale.", "source_urls": ["https://www.nasdaq.com/market-activity/ipos", "https://api.nasdaq.com/api/ipo/calendar?date=2026-10"]}
+{"id": "F10", "use_case": "finance", "task": "When does JPMorgan report Q3?", "search_call": "JPMorgan third quarter 2026 earnings release date", "parameters": "window 7 days", "needs_back": "Date, time of release and call", "freshness": "Week (upcoming)", "answer_key": "Tuesday Oct 13, 2026: results before the open (~7:00 a.m. ET), call at 8:30 a.m. ET.", "source_urls": ["https://www.jpmorganchase.com/ir", "https://web.archive.org/web/20261008095951/https://www.jpmorganchase.com/ir"]}

results.jsonl ADDED Viewed

	@@ -0,0 +1,120 @@

+{"id": "G1", "use_case": "gtm", "provider": "You.com", "returned": "Issues opened Oct 7, just outside the 6 h window", "accuracy_0_2": 0.0}
+{"id": "G1", "use_case": "gtm", "provider": "Tavily", "returned": "No results", "accuracy_0_2": 0.0}
+{"id": "G1", "use_case": "gtm", "provider": "Exa", "returned": "Generic repo issue-list pages, no Kafka issue", "accuracy_0_2": 0.0}
+{"id": "G1", "use_case": "gtm", "provider": "Parallel", "returned": "Undated Kafka-lag tool repos, no issues", "accuracy_0_2": 0.0}
+{"id": "G2", "use_case": "gtm", "provider": "You.com", "returned": "Keyword noise; missed the Oct 8 ATO interview", "accuracy_0_2": 0.0}
+{"id": "G2", "use_case": "gtm", "provider": "Tavily", "returned": "ATO interview with Confluent's VP (Oct 8), in 2 of 3 runs", "accuracy_0_2": 1.33}
+{"id": "G2", "use_case": "gtm", "provider": "Exa", "returned": "ATO interview with Confluent's VP (Oct 8)", "accuracy_0_2": 2.0}
+{"id": "G2", "use_case": "gtm", "provider": "Parallel", "returned": "Undated newsroom and profile pages", "accuracy_0_2": 0.0}
+{"id": "G3", "use_case": "gtm", "provider": "You.com", "returned": "5 in-window apache/iggy issues (Oct 8)", "accuracy_0_2": 2.0}
+{"id": "G3", "use_case": "gtm", "provider": "Tavily", "returned": "Undated Reddit post, rest noise", "accuracy_0_2": 0.0}
+{"id": "G3", "use_case": "gtm", "provider": "Exa", "returned": "Matched \"Apache\" only, no Iggy", "accuracy_0_2": 0.0}
+{"id": "G3", "use_case": "gtm", "provider": "Parallel", "returned": "Evergreen landing pages, nothing dated", "accuracy_0_2": 0.0}
+{"id": "G4", "use_case": "gtm", "provider": "You.com", "returned": "GitHub issues, no release page", "accuracy_0_2": 0.0}
+{"id": "G4", "use_case": "gtm", "provider": "Tavily", "returned": "No results", "accuracy_0_2": 0.0}
+{"id": "G4", "use_case": "gtm", "provider": "Exa", "returned": "Other Apache projects' issue lists", "accuracy_0_2": 0.0}
+{"id": "G4", "use_case": "gtm", "provider": "Parallel", "returned": "server-0.9.0 (Sep 18) with release link", "accuracy_0_2": 2.0}
+{"id": "G5", "use_case": "gtm", "provider": "You.com", "returned": "Right official pages, but both settings cut from the text", "accuracy_0_2": 0.0}
+{"id": "G5", "use_case": "gtm", "provider": "Tavily", "returned": "Confluent docs, both settings described", "accuracy_0_2": 2.0}
+{"id": "G5", "use_case": "gtm", "provider": "Exa", "returned": "Official Kafka pages, only fetch.min.bytes visible", "accuracy_0_2": 0.0}
+{"id": "G5", "use_case": "gtm", "provider": "Parallel", "returned": "max.poll.records named, not described", "accuracy_0_2": 0.0}
+{"id": "G6", "use_case": "gtm", "provider": "You.com", "returned": "GitHub issues, no release page", "accuracy_0_2": 0.0}
+{"id": "G6", "use_case": "gtm", "provider": "Tavily", "returned": "Other projects' release pages", "accuracy_0_2": 0.0}
+{"id": "G6", "use_case": "gtm", "provider": "Exa", "returned": "nats-io issue lists, no release", "accuracy_0_2": 0.0}
+{"id": "G6", "use_case": "gtm", "provider": "Parallel", "returned": "v2.15.1-RC.2 (Oct 6) and v2.15.0 stable", "accuracy_0_2": 2.0}
+{"id": "G7", "use_case": "gtm", "provider": "You.com", "returned": "No active incident; flagged the cached Sep 21 page as stale", "accuracy_0_2": 2.0}
+{"id": "G7", "use_case": "gtm", "provider": "Tavily", "returned": "No results, read as \"none\"", "accuracy_0_2": 2.0}
+{"id": "G7", "use_case": "gtm", "provider": "Exa", "returned": "No results, read as \"none\"", "accuracy_0_2": 2.0}
+{"id": "G7", "use_case": "gtm", "provider": "Parallel", "returned": "No active incident; history page header only", "accuracy_0_2": 2.0}
+{"id": "G8", "use_case": "gtm", "provider": "You.com", "returned": "Manus $500M+ (Oct 8), no lead investor", "accuracy_0_2": 2.0}
+{"id": "G8", "use_case": "gtm", "provider": "Tavily", "returned": "TODAY €2.8M seed (Oct 8), Monid $7.7M", "accuracy_0_2": 2.0}
+{"id": "G8", "use_case": "gtm", "provider": "Exa", "returned": "5 in-window raises (Ampersand, Monid, Nous, Rein)", "accuracy_0_2": 2.0}
+{"id": "G8", "use_case": "gtm", "provider": "Parallel", "returned": "Rein Security $25M (Oct 8)", "accuracy_0_2": 2.0}
+{"id": "G9", "use_case": "gtm", "provider": "You.com", "returned": "Arize, dev.to and Ability.ai articles (Oct 2-6)", "accuracy_0_2": 2.0}
+{"id": "G9", "use_case": "gtm", "provider": "Tavily", "returned": "4 in-window articles (Oct 4-7)", "accuracy_0_2": 2.0}
+{"id": "G9", "use_case": "gtm", "provider": "Exa", "returned": "4 in-window items (Like One, Databricks, arXiv, AgentLensHQ)", "accuracy_0_2": 2.0}
+{"id": "G9", "use_case": "gtm", "provider": "Parallel", "returned": "On-topic, but nothing dated this week", "accuracy_0_2": 0.0}
+{"id": "G10", "use_case": "gtm", "provider": "You.com", "returned": "Medium p99 numbers in window; who ran them not stated", "accuracy_0_2": 1.0}
+{"id": "G10", "use_case": "gtm", "provider": "Tavily", "returned": "Vendor pages, no Redpanda vs Kafka p99", "accuracy_0_2": 0.0}
+{"id": "G10", "use_case": "gtm", "provider": "Exa", "returned": "5 dated benchmarks with authors, no p99 numbers", "accuracy_0_2": 1.0}
+{"id": "G10", "use_case": "gtm", "provider": "Parallel", "returned": "In-window benchmark without p99", "accuracy_0_2": 1.0}
+{"id": "I1", "use_case": "insurance", "provider": "You.com", "returned": "2020-23 Citizens minutes and contracts; active suspension missed", "accuracy_0_2": 0.0}
+{"id": "I1", "use_case": "insurance", "provider": "Tavily", "returned": "No results; suspension missed", "accuracy_0_2": 0.0}
+{"id": "I1", "use_case": "insurance", "provider": "Exa", "returned": "No results; suspension missed", "accuracy_0_2": 0.0}
+{"id": "I1", "use_case": "insurance", "provider": "Parallel", "returned": "Evergreen binding rule + 2024 Helene page; suspension missed", "accuracy_0_2": 0.0}
+{"id": "I2", "use_case": "insurance", "provider": "You.com", "returned": "10 AM CDT advisory (#8, by time): 23.7N 90.2W, 86 mph", "accuracy_0_2": 2.0}
+{"id": "I2", "use_case": "insurance", "provider": "Tavily", "returned": "Stale: newest item Advisory #6A", "accuracy_0_2": 0.0}
+{"id": "I2", "use_case": "insurance", "provider": "Exa", "returned": "Advisory #8, 10 AM CDT, ~85 mph; warning area includes Pensacola", "accuracy_0_2": 2.0}
+{"id": "I2", "use_case": "insurance", "provider": "Parallel", "returned": "10 AM CDT advisory, 86 mph; a stale 8 a.m. advisory in 1 of 3 runs", "accuracy_0_2": 1.33}
+{"id": "I3", "use_case": "insurance", "provider": "You.com", "returned": "Hurricane Warning only, no time; Storm Surge Warning missing", "accuracy_0_2": 1.0}
+{"id": "I3", "use_case": "insurance", "provider": "Tavily", "returned": "Hurricane Warning only; Storm Surge Warning missing", "accuracy_0_2": 1.0}
+{"id": "I3", "use_case": "insurance", "provider": "Exa", "returned": "Both warnings for Escambia, issued 9:53 PM CDT Oct 7", "accuracy_0_2": 2.0}
+{"id": "I3", "use_case": "insurance", "provider": "Parallel", "returned": "Hurricane Warning (issued ~10 p.m. Oct 7); Storm Surge Warning missing", "accuracy_0_2": 1.0}
+{"id": "I4", "use_case": "insurance", "provider": "You.com", "returned": "Stale cached weather pages (NWS block from Aug 5, Fox Weather from Feb 16)", "accuracy_0_2": 0.0}
+{"id": "I4", "use_case": "insurance", "provider": "Tavily", "returned": "Averages and forecasts; current 13 mph ENE (16:00Z) on repeats 2-3 only", "accuracy_0_2": 1.33}
+{"id": "I4", "use_case": "insurance", "provider": "Exa", "returned": "13 mph NE at 11:00 AM CDT (model estimate)", "accuracy_0_2": 2.0}
+{"id": "I4", "use_case": "insurance", "provider": "Parallel", "returned": "Forecasts only; newest station report ~5 h old", "accuracy_0_2": 0.0}
+{"id": "I5", "use_case": "insurance", "provider": "You.com", "returned": "No mandatory order; voluntary coastal evacuation (Oct 7)", "accuracy_0_2": 2.0}
+{"id": "I5", "use_case": "insurance", "provider": "Tavily", "returned": "No mandatory order; voluntary evacuation in effect", "accuracy_0_2": 2.0}
+{"id": "I5", "use_case": "insurance", "provider": "Exa", "returned": "Explicitly \"no mandatory evacuation\"; voluntary in effect", "accuracy_0_2": 2.0}
+{"id": "I5", "use_case": "insurance", "provider": "Parallel", "returned": "No mandatory order; voluntary recommendation (Oct 7)", "accuracy_0_2": 2.0}
+{"id": "I6", "use_case": "insurance", "provider": "You.com", "returned": "None yet; only Florida's state of emergency", "accuracy_0_2": 2.0}
+{"id": "I6", "use_case": "insurance", "provider": "Tavily", "returned": "None yet; off-target results", "accuracy_0_2": 2.0}
+{"id": "I6", "use_case": "insurance", "provider": "Exa", "returned": "None yet; no 2020 declaration surfaced", "accuracy_0_2": 2.0}
+{"id": "I6", "use_case": "insurance", "provider": "Parallel", "returned": "None yet; FEMA press page (Oct 7) shows no Isaias declaration", "accuracy_0_2": 2.0}
+{"id": "I7", "use_case": "insurance", "provider": "You.com", "returned": "None; no regulator order in results", "accuracy_0_2": 2.0}
+{"id": "I7", "use_case": "insurance", "provider": "Tavily", "returned": "No results (correct none by default)", "accuracy_0_2": 2.0}
+{"id": "I7", "use_case": "insurance", "provider": "Exa", "returned": "None; found the Oct 7 FLOIR notice and ALDOI release and did not call them orders", "accuracy_0_2": 2.0}
+{"id": "I7", "use_case": "insurance", "provider": "Parallel", "returned": "None; old Helene order labelled as old", "accuracy_0_2": 2.0}
+{"id": "I8", "use_case": "insurance", "provider": "You.com", "returned": "FL EO 26-202 (Oct 6); AL named without its date", "accuracy_0_2": 1.0}
+{"id": "I8", "use_case": "insurance", "provider": "Tavily", "returned": "FL (Oct 6) + AL (Oct 7)", "accuracy_0_2": 2.0}
+{"id": "I8", "use_case": "insurance", "provider": "Exa", "returned": "FL EO 26-202 (Oct 6) only; AL missing", "accuracy_0_2": 1.0}
+{"id": "I8", "use_case": "insurance", "provider": "Parallel", "returned": "AL (Oct 7) + FL EO 26-202 (Oct 6)", "accuracy_0_2": 2.0}
+{"id": "I9", "use_case": "insurance", "provider": "You.com", "returned": "None: a 1998 SPC paper and the reports index, no storm data", "accuracy_0_2": 2.0}
+{"id": "I9", "use_case": "insurance", "provider": "Tavily", "returned": "No results, read as \"none\"", "accuracy_0_2": 2.0}
+{"id": "I9", "use_case": "insurance", "provider": "Exa", "returned": "No results, read as \"none\"", "accuracy_0_2": 2.0}
+{"id": "I9", "use_case": "insurance", "provider": "Parallel", "returned": "No results, read as \"none\"", "accuracy_0_2": 2.0}
+{"id": "I10", "use_case": "insurance", "provider": "You.com", "returned": "Only 26V179 (2024 RAV4 Hybrid label); 24V911000 missed", "accuracy_0_2": 1.0}
+{"id": "I10", "use_case": "insurance", "provider": "Tavily", "returned": "25V595 + 25V744 + 26V179; 24V911000 missed", "accuracy_0_2": 1.0}
+{"id": "I10", "use_case": "insurance", "provider": "Exa", "returned": "2 results: a Lexus recall excerpt and a Nissan page; no RAV4 recall", "accuracy_0_2": 0.0}
+{"id": "I10", "use_case": "insurance", "provider": "Parallel", "returned": "25V744 + 26V179; 24V911000 missed", "accuracy_0_2": 1.0}
+{"id": "F1", "use_case": "finance", "provider": "You.com", "returned": "$235.72 dated Oct 8, no intraday time (Robinhood)", "accuracy_0_2": 2.0}
+{"id": "F1", "use_case": "finance", "provider": "Tavily", "returned": "$235.22 at 12:05 PM EDT (Yahoo); only the Oct 7 close in 1 of 3 runs", "accuracy_0_2": 1.33}
+{"id": "F1", "use_case": "finance", "provider": "Exa", "returned": "Stale: the Oct 7 close of $237.47", "accuracy_0_2": 0.0}
+{"id": "F1", "use_case": "finance", "provider": "Parallel", "returned": "$235.25 at 12:21 PM EDT (StockAnalysis)", "accuracy_0_2": 2.0}
+{"id": "F2", "use_case": "finance", "provider": "You.com", "returned": "~5.27% (Trading Economics; CNBC 5.284% at 12:22 PM)", "accuracy_0_2": 2.0}
+{"id": "F2", "use_case": "finance", "provider": "Tavily", "returned": "No current yield: old Binance posts (4.33%, March); morning highs on repeats", "accuracy_0_2": 0.0}
+{"id": "F2", "use_case": "finance", "provider": "Exa", "returned": "Stale: 5.322% from 5:19 AM, a morning high outside ±3 bp", "accuracy_0_2": 0.0}
+{"id": "F2", "use_case": "finance", "provider": "Parallel", "returned": "5.273% (CNBC); 5.28% (Trading Economics)", "accuracy_0_2": 2.0}
+{"id": "F3", "use_case": "finance", "provider": "You.com", "returned": "$81,307 (CoinGecko), within 0.2% of spot", "accuracy_0_2": 2.0}
+{"id": "F3", "use_case": "finance", "provider": "Tavily", "returned": "$81,028 at 15:57 UTC (Yahoo)", "accuracy_0_2": 2.0}
+{"id": "F3", "use_case": "finance", "provider": "Exa", "returned": "No point quote: \"below $82,000\" and older $82,650-82,970", "accuracy_0_2": 0.0}
+{"id": "F3", "use_case": "finance", "provider": "Parallel", "returned": "$81,063 (CoinMarketCap); $81,248 (CoinGecko)", "accuracy_0_2": 2.0}
+{"id": "F4", "use_case": "finance", "provider": "You.com", "returned": "Stale: $92.63 at 8:18 AM ET, +1.7% vs the key", "accuracy_0_2": 0.0}
+{"id": "F4", "use_case": "finance", "provider": "Tavily", "returned": "$91.58, an Asian-session figure in tolerance by luck; $92.90 (11:52 AM) in 1 of 3 runs", "accuracy_0_2": 1.33}
+{"id": "F4", "use_case": "finance", "provider": "Exa", "returned": "Stale: $92.29 at 10:43 AM EDT, +1.4%", "accuracy_0_2": 0.0}
+{"id": "F4", "use_case": "finance", "provider": "Parallel", "returned": "$90.91 (OilPrice.com, 10-min delayed); stale quotes in 2 of 3 runs", "accuracy_0_2": 0.67}
+{"id": "F5", "use_case": "finance", "provider": "You.com", "returned": "None: 2 irrelevant dol.gov pages (2009-10 archive, a FOIA PDF)", "accuracy_0_2": 0.0}
+{"id": "F5", "use_case": "finance", "provider": "Tavily", "returned": "No results (dol.gov filter)", "accuracy_0_2": 0.0}
+{"id": "F5", "use_case": "finance", "provider": "Exa", "returned": "No results (dol.gov filter)", "accuracy_0_2": 0.0}
+{"id": "F5", "use_case": "finance", "provider": "Parallel", "returned": "197,000 for the week ending Oct 3 (DOL PDF, released 8:30 a.m. Oct 8)", "accuracy_0_2": 2.0}
+{"id": "F6", "use_case": "finance", "provider": "You.com", "returned": "Sep 15-16 minutes, released Oct 7, federalreserve.gov link, 2 key points", "accuracy_0_2": 2.0}
+{"id": "F6", "use_case": "finance", "provider": "Tavily", "returned": "No results (federalreserve.gov filter)", "accuracy_0_2": 0.0}
+{"id": "F6", "use_case": "finance", "provider": "Exa", "returned": "Sep 15-16 minutes, released Oct 7, with link; 1 key point", "accuracy_0_2": 2.0}
+{"id": "F6", "use_case": "finance", "provider": "Parallel", "returned": "Sep 15-16 minutes, released Oct 7, press-release link; no key points", "accuracy_0_2": 2.0}
+{"id": "F7", "use_case": "finance", "provider": "You.com", "returned": "Viatris-Pacira $1.65B (Oct 8) + borderline RXO, Hormel, Sunrise; old Roche deal excluded", "accuracy_0_2": 2.0}
+{"id": "F7", "use_case": "finance", "provider": "Tavily", "returned": "Schneider-PTC (Oct 5, out of window) shown as new; RXO borderline", "accuracy_0_2": 0.0}
+{"id": "F7", "use_case": "finance", "provider": "Exa", "returned": "Wittington-Boots $8.9B (Oct 7) + Viatris-Pacira $1.65B (Oct 8)", "accuracy_0_2": 2.0}
+{"id": "F7", "use_case": "finance", "provider": "Parallel", "returned": "None in window (July and Sept deals); in-window deals on repeats 2-3", "accuracy_0_2": 1.0}
+{"id": "F8", "use_case": "finance", "provider": "You.com", "returned": "Oct 6 8-K, Items 5.02/7.01/9.01, EDGAR index link; Stiles named incoming CEO", "accuracy_0_2": 2.0}
+{"id": "F8", "use_case": "finance", "provider": "Tavily", "returned": "No results (sec.gov filter)", "accuracy_0_2": 0.0}
+{"id": "F8", "use_case": "finance", "provider": "Exa", "returned": "Other issuers' 8-Ks; no Stryker filing", "accuracy_0_2": 0.0}
+{"id": "F8", "use_case": "finance", "provider": "Parallel", "returned": "Oct 6 8-K: Lobo to Executive Chair, Stiles CEO from Jan 1, 2027", "accuracy_0_2": 2.0}
+{"id": "F9", "use_case": "finance", "provider": "You.com", "returned": "$5.60 per ADS, CHWM, Nasdaq Capital Market, priced Sep 30/Oct 1", "accuracy_0_2": 2.0}
+{"id": "F9", "use_case": "finance", "provider": "Tavily", "returned": "$5.60, CHWM, Nasdaq, Oct 1", "accuracy_0_2": 2.0}
+{"id": "F9", "use_case": "finance", "provider": "Exa", "returned": "$5.60 per ADS, Nasdaq Capital Market, Oct 1 (424B1 confirms)", "accuracy_0_2": 2.0}
+{"id": "F9", "use_case": "finance", "provider": "Parallel", "returned": "$5.60, CHWM, Nasdaq Capital Market, Oct 1; flagged stale $5.00 terms", "accuracy_0_2": 2.0}
+{"id": "F10", "use_case": "finance", "provider": "You.com", "returned": "Tue Oct 13, release ~7:00 a.m. ET, call 8:30 a.m.", "accuracy_0_2": 2.0}
+{"id": "F10", "use_case": "finance", "provider": "Tavily", "returned": "Oct 13, ~7:00 release, 8:30 call; other companies' dates in 1 of 3 runs", "accuracy_0_2": 1.33}
+{"id": "F10", "use_case": "finance", "provider": "Exa", "returned": "Oct 13, before the open; call 8:30 a.m. ET", "accuracy_0_2": 2.0}
+{"id": "F10", "use_case": "finance", "provider": "Parallel", "returned": "Oct 13, results ~7:00 a.m. ET, call 8:30 a.m.", "accuracy_0_2": 2.0}

scorecards.jsonl ADDED Viewed

	@@ -0,0 +1,136 @@

+{"scope": "overall", "metric": "Accuracy score (% of max)", "provider": "You.com", "value": "63%"}
+{"scope": "overall", "metric": "Accuracy score (% of max)", "provider": "Tavily", "value": "51%"}
+{"scope": "overall", "metric": "Accuracy score (% of max)", "provider": "Exa", "value": "53%"}
+{"scope": "overall", "metric": "Accuracy score (% of max)", "provider": "Parallel", "value": "67%"}
+{"scope": "overall", "metric": "Queries fully correct in all 3 repeats", "provider": "You.com", "value": "57%"}
+{"scope": "overall", "metric": "Queries fully correct in all 3 repeats", "provider": "Tavily", "value": "37%"}
+{"scope": "overall", "metric": "Queries fully correct in all 3 repeats", "provider": "Exa", "value": "50%"}
+{"scope": "overall", "metric": "Queries fully correct in all 3 repeats", "provider": "Parallel", "value": "57%"}
+{"scope": "overall", "metric": "Precision@5 (on-topic share of top 5)", "provider": "You.com", "value": "37%"}
+{"scope": "overall", "metric": "Precision@5 (on-topic share of top 5)", "provider": "Tavily", "value": "26%"}
+{"scope": "overall", "metric": "Precision@5 (on-topic share of top 5)", "provider": "Exa", "value": "40%"}
+{"scope": "overall", "metric": "Precision@5 (on-topic share of top 5)", "provider": "Parallel", "value": "26%"}
+{"scope": "overall", "metric": "Results dated inside the query's window", "provider": "You.com", "value": "59%"}
+{"scope": "overall", "metric": "Results dated inside the query's window", "provider": "Tavily", "value": "37%"}
+{"scope": "overall", "metric": "Results dated inside the query's window", "provider": "Exa", "value": "44%"}
+{"scope": "overall", "metric": "Results dated inside the query's window", "provider": "Parallel", "value": "29%"}
+{"scope": "overall", "metric": "Newest relevant result, median age", "provider": "You.com", "value": "4.1 h"}
+{"scope": "overall", "metric": "Newest relevant result, median age", "provider": "Tavily", "value": "11.9 h"}
+{"scope": "overall", "metric": "Newest relevant result, median age", "provider": "Exa", "value": "8.2 h"}
+{"scope": "overall", "metric": "Newest relevant result, median age", "provider": "Parallel", "value": "16.6 h"}
+{"scope": "overall", "metric": "Empty responses (of 90)", "provider": "You.com", "value": "0"}
+{"scope": "overall", "metric": "Empty responses (of 90)", "provider": "Tavily", "value": "27"}
+{"scope": "overall", "metric": "Empty responses (of 90)", "provider": "Exa", "value": "12"}
+{"scope": "overall", "metric": "Empty responses (of 90)", "provider": "Parallel", "value": "3"}
+{"scope": "overall", "metric": "Failed calls: errors, timeouts (of 90)", "provider": "You.com", "value": "0"}
+{"scope": "overall", "metric": "Failed calls: errors, timeouts (of 90)", "provider": "Tavily", "value": "0"}
+{"scope": "overall", "metric": "Failed calls: errors, timeouts (of 90)", "provider": "Exa", "value": "0"}
+{"scope": "overall", "metric": "Failed calls: errors, timeouts (of 90)", "provider": "Parallel", "value": "0"}
+{"scope": "overall", "metric": "Latency p50 / p95, client (ms)", "provider": "You.com", "value": "593 / 804"}
+{"scope": "overall", "metric": "Latency p50 / p95, client (ms)", "provider": "Tavily", "value": "1,864 / 3,305"}
+{"scope": "overall", "metric": "Latency p50 / p95, client (ms)", "provider": "Exa", "value": "1,106 / 3,997"}
+{"scope": "overall", "metric": "Latency p50 / p95, client (ms)", "provider": "Parallel", "value": "2,038 / 4,117"}
+{"scope": "overall", "metric": "Latency p50, server-reported (ms)", "provider": "You.com", "value": "379"}
+{"scope": "overall", "metric": "Latency p50, server-reported (ms)", "provider": "Tavily", "value": "1,650"}
+{"scope": "overall", "metric": "Latency p50, server-reported (ms)", "provider": "Exa", "value": "831"}
+{"scope": "overall", "metric": "Latency p50, server-reported (ms)", "provider": "Parallel", "value": "n/a"}
+{"scope": "overall", "metric": "Search cost per correct answer", "provider": "You.com", "value": "$0.026"}
+{"scope": "overall", "metric": "Search cost per correct answer", "provider": "Tavily", "value": "$0.016"}
+{"scope": "overall", "metric": "Search cost per correct answer", "provider": "Exa", "value": "$0.012"}
+{"scope": "overall", "metric": "Search cost per correct answer", "provider": "Parallel", "value": "$0.008"}
+{"scope": "gtm", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "You.com", "value": "381 / 627 · 512 / 751 (fastest)"}
+{"scope": "gtm", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Tavily", "value": "2,670 / 2,843 · 3,295 / 3,425"}
+{"scope": "gtm", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Exa", "value": "812 / 1,102 · 3,178 / 3,403"}
+{"scope": "gtm", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Parallel", "value": "n/a / 2,268 · n/a / 4,268"}
+{"scope": "gtm", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "You.com", "value": "10 of 10; slowest 781 ms"}
+{"scope": "gtm", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Tavily", "value": "0 of 10; slowest 3,473 ms"}
+{"scope": "gtm", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Exa", "value": "3 of 10; slowest 3,419 ms"}
+{"scope": "gtm", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Parallel", "value": "1 of 10; slowest 4,289 ms"}
+{"scope": "gtm", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "You.com", "value": "45%"}
+{"scope": "gtm", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Tavily", "value": "47%"}
+{"scope": "gtm", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Exa", "value": "45%"}
+{"scope": "gtm", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Parallel", "value": "45%"}
+{"scope": "gtm", "metric": "Freshness: % in window", "provider": "You.com", "value": "75%"}
+{"scope": "gtm", "metric": "Freshness: % in window", "provider": "Tavily", "value": "47%"}
+{"scope": "gtm", "metric": "Freshness: % in window", "provider": "Exa", "value": "58%"}
+{"scope": "gtm", "metric": "Freshness: % in window", "provider": "Parallel", "value": "28%"}
+{"scope": "gtm", "metric": "Freshness: newest relevant result (h)", "provider": "You.com", "value": "4.1"}
+{"scope": "gtm", "metric": "Freshness: newest relevant result (h)", "provider": "Tavily", "value": "39.9"}
+{"scope": "gtm", "metric": "Freshness: newest relevant result (h)", "provider": "Exa", "value": "15.9"}
+{"scope": "gtm", "metric": "Freshness: newest relevant result (h)", "provider": "Parallel", "value": "399.9"}
+{"scope": "gtm", "metric": "Content in one call", "provider": "You.com", "value": "100%"}
+{"scope": "gtm", "metric": "Content in one call", "provider": "Tavily", "value": "70%"}
+{"scope": "gtm", "metric": "Content in one call", "provider": "Exa", "value": "90%"}
+{"scope": "gtm", "metric": "Content in one call", "provider": "Parallel", "value": "100%"}
+{"scope": "gtm", "metric": "Tokens, top-5 results: per query / per correct", "provider": "You.com", "value": "95,707 / 239,254"}
+{"scope": "gtm", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Tavily", "value": "20,199 / 42,276"}
+{"scope": "gtm", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Exa", "value": "10,563 / 26,409"}
+{"scope": "gtm", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Parallel", "value": "1,814 / 4,887"}
+{"scope": "gtm", "metric": "Search cost: per query / per correct ($)", "provider": "You.com", "value": "0.0150 / 0.0375"}
+{"scope": "gtm", "metric": "Search cost: per query / per correct ($)", "provider": "Tavily", "value": "0.0080 / 0.0160"}
+{"scope": "gtm", "metric": "Search cost: per query / per correct ($)", "provider": "Exa", "value": "0.0063 / 0.0158"}
+{"scope": "gtm", "metric": "Search cost: per query / per correct ($)", "provider": "Parallel", "value": "0.0050 / 0.0125"}
+{"scope": "insurance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "You.com", "value": "379 / 589 · 496 / 772 (fastest)"}
+{"scope": "insurance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Tavily", "value": "1,650 / 1,864 · 2,959 / 3,128"}
+{"scope": "insurance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Exa", "value": "856 / 1,093 · 4,194 / 4,448"}
+{"scope": "insurance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Parallel", "value": "n/a / 2,038 · n/a / 3,450"}
+{"scope": "insurance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "You.com", "value": "10 of 10; slowest 865 ms"}
+{"scope": "insurance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Tavily", "value": "1 of 10; slowest 3,230 ms"}
+{"scope": "insurance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Exa", "value": "3 of 10; slowest 4,593 ms"}
+{"scope": "insurance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Parallel", "value": "1 of 10; slowest 3,963 ms"}
+{"scope": "insurance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "You.com", "value": "65%"}
+{"scope": "insurance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Tavily", "value": "67%"}
+{"scope": "insurance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Exa", "value": "75%"}
+{"scope": "insurance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Parallel", "value": "67%"}
+{"scope": "insurance", "metric": "Freshness: % in window", "provider": "You.com", "value": "56%"}
+{"scope": "insurance", "metric": "Freshness: % in window", "provider": "Tavily", "value": "34%"}
+{"scope": "insurance", "metric": "Freshness: % in window", "provider": "Exa", "value": "31%"}
+{"scope": "insurance", "metric": "Freshness: % in window", "provider": "Parallel", "value": "31%"}
+{"scope": "insurance", "metric": "Freshness: newest relevant result (h)", "provider": "You.com", "value": "2.3"}
+{"scope": "insurance", "metric": "Freshness: newest relevant result (h)", "provider": "Tavily", "value": "11.9"}
+{"scope": "insurance", "metric": "Freshness: newest relevant result (h)", "provider": "Exa", "value": "7.1"}
+{"scope": "insurance", "metric": "Freshness: newest relevant result (h)", "provider": "Parallel", "value": "16.3"}
+{"scope": "insurance", "metric": "Content in one call", "provider": "You.com", "value": "100%"}
+{"scope": "insurance", "metric": "Content in one call", "provider": "Tavily", "value": "70%"}
+{"scope": "insurance", "metric": "Content in one call", "provider": "Exa", "value": "80%"}
+{"scope": "insurance", "metric": "Content in one call", "provider": "Parallel", "value": "90%"}
+{"scope": "insurance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "You.com", "value": "14,524 / 29,048"}
+{"scope": "insurance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Tavily", "value": "13,986 / 28,383"}
+{"scope": "insurance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Exa", "value": "4,731 / 6,759"}
+{"scope": "insurance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Parallel", "value": "1,418 / 2,352"}
+{"scope": "insurance", "metric": "Search cost: per query / per correct ($)", "provider": "You.com", "value": "0.0150 / 0.0300"}
+{"scope": "insurance", "metric": "Search cost: per query / per correct ($)", "provider": "Tavily", "value": "0.0080 / 0.0160"}
+{"scope": "insurance", "metric": "Search cost: per query / per correct ($)", "provider": "Exa", "value": "0.0056 / 0.0080"}
+{"scope": "insurance", "metric": "Search cost: per query / per correct ($)", "provider": "Parallel", "value": "0.0050 / 0.0083"}
+{"scope": "finance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "You.com", "value": "372 / 592 · 422 / 776 (fastest)"}
+{"scope": "finance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Tavily", "value": "1,465 / 1,642 · 1,699 / 1,889"}
+{"scope": "finance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Exa", "value": "870 / 1,106 · 2,743 / 3,008"}
+{"scope": "finance", "metric": "Search-only latency p50 · p95, server / client (ms)", "provider": "Parallel", "value": "n/a / 1,972 · n/a / 3,511"}
+{"scope": "finance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "You.com", "value": "10 of 10; slowest 823 ms"}
+{"scope": "finance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Tavily", "value": "0 of 10; slowest 1,997 ms"}
+{"scope": "finance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Exa", "value": "3 of 10; slowest 3,170 ms"}
+{"scope": "finance", "metric": "Search-only calls ≤1 s, client (of 10); slowest call", "provider": "Parallel", "value": "0 of 10; slowest 3,847 ms"}
+{"scope": "finance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "You.com", "value": "80%"}
+{"scope": "finance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Tavily", "value": "40%"}
+{"scope": "finance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Exa", "value": "40%"}
+{"scope": "finance", "metric": "Accuracy score (% of max, 10 queries × 3 repeats)", "provider": "Parallel", "value": "88%"}
+{"scope": "finance", "metric": "Freshness: % in window", "provider": "You.com", "value": "45%"}
+{"scope": "finance", "metric": "Freshness: % in window", "provider": "Tavily", "value": "31%"}
+{"scope": "finance", "metric": "Freshness: % in window", "provider": "Exa", "value": "44%"}
+{"scope": "finance", "metric": "Freshness: % in window", "provider": "Parallel", "value": "29%"}
+{"scope": "finance", "metric": "Freshness: newest relevant result (h)", "provider": "You.com", "value": "6.4"}
+{"scope": "finance", "metric": "Freshness: newest relevant result (h)", "provider": "Tavily", "value": "5.6"}
+{"scope": "finance", "metric": "Freshness: newest relevant result (h)", "provider": "Exa", "value": "8.2"}
+{"scope": "finance", "metric": "Freshness: newest relevant result (h)", "provider": "Parallel", "value": "16.6"}
+{"scope": "finance", "metric": "Content in one call", "provider": "You.com", "value": "90%"}
+{"scope": "finance", "metric": "Content in one call", "provider": "Tavily", "value": "70%"}
+{"scope": "finance", "metric": "Content in one call", "provider": "Exa", "value": "90%"}
+{"scope": "finance", "metric": "Content in one call", "provider": "Parallel", "value": "100%"}
+{"scope": "finance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "You.com", "value": "13,146 / 16,433"}
+{"scope": "finance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Tavily", "value": "11,737 / 25,644"}
+{"scope": "finance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Exa", "value": "9,785 / 23,613"}
+{"scope": "finance", "metric": "Tokens, top-5 results: per query / per correct", "provider": "Parallel", "value": "1,793 / 1,999"}
+{"scope": "finance", "metric": "Search cost: per query / per correct ($)", "provider": "You.com", "value": "0.0150 / 0.0188"}
+{"scope": "finance", "metric": "Search cost: per query / per correct ($)", "provider": "Tavily", "value": "0.0080 / 0.0160"}
+{"scope": "finance", "metric": "Search cost: per query / per correct ($)", "provider": "Exa", "value": "0.0063 / 0.0158"}
+{"scope": "finance", "metric": "Search cost: per query / per correct ($)", "provider": "Parallel", "value": "0.0050 / 0.0056"}