METR Q4 2026 TH2 - Judge Run (Newer Models' Tasks): leaderboard
METR's Q4 2026 capability update (20 November 2026): mean score from 0 to 1 on one task of TH2, METR's new long-task suite. The agent grades another agent's attempt at building an evaluation task, using a rubric, and is scored on agreement with METR's human grader. This half uses tasks built by models released from 9 June 2026, where every judge agrees with humans less. Higher is better.
Metric: Mean Score (0-1). Source: metr.org. Saturation forecast: Around May 2027. 15 models tracked. Updated 21 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 | 0.4586 | #9 |
| 2 | Claude Fable 5.1 | 0.35 | #20 |
| 3 | GPT-6 Astra | 0.2653 | #10 |
| 4 | Claude Opus 5 | 0.1769 | #26 |
| 5 | Claude Opus 4.7 | 0.1476 | #85 |
| 6 | Claude Fable 5 | 0.1461 | #33 |
| 7 | Claude Opus 4.5 | 0.1095 | #160 |
| 8 | Claude Opus 4.6 | 0.0919 | #113 |
| 9 | GPT-5.6 Sol | 0.0666 | #37 |
| 10 | GPT-5.4 | 0.0551 | #135 |
| 11 | GPT-5 | 0.0387 | #178 |
| 12 | O1 | 0.0367 | #403 |
| 13 | GPT-5.5 | 0.0288 | #54 |
| 14 | Claude Opus 4.8 | 0.0263 | #65 |
| 15 | GPT-4o | 0 | #751 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=metr-q4-2026-th2-judge-run-newer-models-tasks · How It Works · Data refreshed daily, snapshot 2026-11-21.