Skip to main content
ModelScale

Provenance

Cited sources

Every number on this site is reproduced from one of the fetches below, in the unit its source declared. This page is the whole list, with the time each one was last read.

What each source supplies

BenchLM
Catalog membership, published rankings, benchmark task results, list pricing, measured speed, pairwise comparisons, and lifecycle notices.
LiveBench
Per-task contamination-free benchmark scores and the cost table published alongside them.
OpenRouter
A hosted price fallback for models the primary source lists without one, and per-endpoint throughput observations.

Fetches in the current dataset

11 sources · dataset generated 2026-09-17 17:17 UTC.

Fetch state key: freshpartial

benchlm:modelsfreshWhich models exist, their identity, limits, and access mode.Fetched 2026-09-17 17:17 UTCSource updated September 17, 2026486 recordshttps://benchlm.ai/data/models.json
benchlm:leaderboardfreshPublished overall rank and the seven capability scores.Fetched 2026-09-17 17:17 UTCSource updated September 17, 2026228 recordshttps://benchlm.ai/data/leaderboard.json
benchlm:benchmarksfreshIndividual benchmark task results, grouped by category.Fetched 2026-09-17 17:17 UTCSource updated September 17, 2026439 recordshttps://benchlm.ai/data/benchmarks.json
benchlm:pricingpartialList token prices: input, output, and cache rates where published.Fetched 2026-09-17 17:17 UTCSource updated September 17, 2026424 recordshttps://benchlm.ai/data/pricing.json
benchlm:speedfreshMeasured time to first token and output throughput.Fetched 2026-09-17 17:17 UTCSource updated 2026-09-15138 recordshttps://benchlm.ai/data/speed.json
benchlm:comparisonsfreshPairwise category deltas used by the compare page.Fetched 2026-09-17 17:17 UTCSource updated September 17, 2026400 recordshttps://benchlm.ai/data/comparisons.json
benchlm:lifecyclefreshRelease and retirement notices with their effective dates.Fetched 2026-09-17 17:17 UTCSource updated 2026-08-1152 recordshttps://benchlm.ai/api/data/ai-model-changes
livebench:tablefreshPer-task LiveBench scores for the models it covers.Fetched 2026-09-17 17:17 UTC58 recordshttps://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv
livebench:costfreshThe cost figures LiveBench publishes beside those scores.Fetched 2026-09-17 17:17 UTC58 recordshttps://raw.githubusercontent.com/LiveBench/new-livebench/main/public/cost_2026_06_25.csv
openrouter:pricingfreshHosted token prices, used only where the primary source publishes none.Fetched 2026-09-17 17:17 UTC444 recordshttps://openrouter.ai/api/v1/models
openrouter:endpointsfreshPer-endpoint throughput observations for hosted deployments.Fetched 2026-09-17 17:17 UTC11 recordshttps://openrouter.ai/api/v1/models/{id}/endpoints

How this site treats evidence

A value is shown exactly as its source published it, in that source’s declared unit — nothing is converted, interpolated, or averaged across providers. Where a source returned no value, the site prints “Unavailable” and records when the fetch was attempted; it never substitutes a zero, and a missing measurement is never ranked as if it were a low one. Where a fetch failed and a previous value is still on file, that value is shown with a “Fetch failed” badge and is never used to compute a cost estimate. Every figure carries an evidence dot linking to the fetch it came from, so the claim on this page can be checked one number at a time; the key beside each table says which states that section contains.

Attribution

Evidence values belong to their cited sources and are reproduced here for comparison. Provider logos come from Lobe Icons (MIT), and each remains the trademark of the provider it identifies.