Claim · reasoning

GPT-5 Nano (2025-08-07)on GPQA Diamond (Epoch's own run)

48.5 to 69.4 the range across every scoreable measurement we hold
21.0 points between the highest and the lowest
57.6 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
69.4 / 100
as published: 69.4%
INDEPENDENT
counts toward the ranking
effort high Epoch AI
2025-10-30T16:20:27.447Z
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
67.4 / 100
as published: 67.4%
INDEPENDENT
counts toward the ranking
effort medium Epoch AI
2025-08-07T19:34:29.562Z
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
57.6 / 100
as published: 57.6%
INDEPENDENT
counts toward the ranking
effort low Epoch AI
2026-07-13T17:28:21.000Z
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
48.5 / 100
as published: 48.5%
INDEPENDENT
counts toward the ranking
effort minimal Epoch AI
2026-07-13T17:28:09.000Z
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.