Claim · coding

DeepSeek V4 Flashon ALE-Bench

9.3 to 19.4 the range across every scoreable measurement we hold
10.1 points between the highest and the lowest
9.3 the number the board uses: the lower median, a real row with its own provenance
2 declared configurations, from 1 runner

All 2 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

2 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
19.4 / 100
as published: 678
 INDEPENDENT
counts toward the ranking
 effort high ALE-Bench (Sakana AI with AtCoder)
via Epoch AI
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
9.3 / 100
as published: 325
 INDEPENDENT
counts toward the ranking
 effort none ALE-Bench (Sakana AI with AtCoder)
via Epoch AI
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.