Claim · reasoning

Claude Opus 4.6on FrontierMath

38.3 to 40.7 the range across every scoreable measurement we hold
2.4 points between the highest and the lowest
39.7 the number the board uses: the lower median, a real row with its own provenance
4 declared configurations, from 1 runner

All 4 scoreable measurements here come from ONE runner, so this is not a reconciliation and is not presented as one. What it shows is how far a single runner's number moves when the model is run differently.

Every measurement

4 rows. Best first. the board takes the LOWER MEDIAN of the scoreable measurements, which is a real row with its own interval and provenance rather than an interpolated average. a vendor self-report is shown beside the independent numbers and is never one of them; it cannot move a rank.

ScoreProvenanceConfigurationWho ran itWhere to check it
40.7 / 100
as published: 40.7%
INDEPENDENT
counts toward the ranking
effort max Epoch AI
2026-02-12T17:17:42.489Z
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
40.0 / 100
as published: 40.0%
INDEPENDENT
counts toward the ranking
effort 32k Epoch AI
2026-02-06T12:46:04.911Z
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
39.7 / 100
as published: 39.7%
INDEPENDENT
counts toward the ranking
effort 64k Epoch AI
2026-02-06T15:39:35.993Z
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
38.3 / 100
as published: 38.3%
INDEPENDENT
counts toward the ranking
not stated by the source Epoch AI
2026-02-05T18:49:51.181Z
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
← The Forge Rankings

Snapshot 2026-08-30-62e4e85f043b, manifest dc46131315046f1e, as of 2026-08-30. where only one party has measured a model on a benchmark, this page says so rather than presenting a single number as a reconciliation.