Claim · reasoning

Gemini 3.6 Flashon ARC-AGI-2

2.6 to 60.4 the range across every scoreable measurement we hold
57.8 points between the highest and the lowest
30.4 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
60.4 / 100
as published: 60.4%
INDEPENDENT
counts toward the ranking
gemini-3.6-flash
effort high
ARC Prize
via Epoch AI
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
50.4 / 100
as published: 50.4%
INDEPENDENT
counts toward the ranking
gemini-3.6-flash
effort medium
ARC Prize
via Epoch AI
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
30.4 / 100
as published: 30.4%
INDEPENDENT
counts toward the ranking
gemini-3.6-flash
effort low
ARC Prize
via Epoch AI
CC-BY-4.0
NO_URL2026-08-30
benchmark leaderboard
the leaderboard for this benchmark, not a link to this row
2.6 / 100
as published: 2.6%
INDEPENDENT
counts toward the ranking
gemini-3.6-flash
effort minimal
ARC Prize
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.