BenchAtlas

Sources / livebench

LiveBench

healthydefault: independent

livebench.ai · trust rank 85 · adapter livebench

Observations
291
append-only score facts
Raw snapshots
54
immutable fetched payloads
Last success
2026-09-12 07:00Z
end of last succeeded/partial run
Freshness
1.3 h
hours since last success

License & attribution

The LiveBench repository is Apache-2.0; the site CSVs themselves are unlabeled and are surfaced as facts with attribution and a link. Newest questions are deliberately withheld under their contamination policy.

What we ingest

Per-release table CSV paired with its same-version categories JSON; task scores average to category means (their method). Each dated release is a separate benchmark version, and inference-config slugs stay separate evaluated systems.

Ingestion runs

last 10 of the run history
StartedDurationStatus
2026-09-12 07:00Z2.0 spartial
2026-09-11 07:00Z2.6 spartial
2026-09-10 07:00Z2.0 spartial
2026-09-09 07:00Z1.7 spartial
2026-09-08 07:00Z1.9 spartial
2026-09-07 07:00Z1.8 spartial
2026-09-06 07:00Z1.4 spartial
2026-09-05 07:00Z1.7 spartial
2026-09-04 07:00Z1.8 spartial
2026-09-03 07:00Z2.1 spartial

partial = run finished but some rows are blocked on identity review — nothing is published until every referenced model/agent/benchmark is resolved.

Raw snapshots

latest 5
FetchedURLSHA-256
2026-09-11 07:00Zhttps://livebench.ai/cost_2026_06_25.csv45a4026f169a
2026-09-11 07:00Zhttps://livebench.ai/table_2026_06_25.csv186ca65e7bb3
2026-09-08 07:00Zhttps://livebench.ai/cost_2026_06_25.csvc8c24e6f5091
2026-09-08 07:00Zhttps://livebench.ai/table_2026_06_25.csva348f9388138
2026-09-07 07:00Zhttps://livebench.ai/cost_2026_06_25.csvd121c05b440f

Snapshots are immutable raw payloads — every published number traces back to one. Re-fetching an identical payload bumps the re-fetch time instead of storing a duplicate. Scores from this source appear on benchmark profiles and model pages with this source linked.