METR Q4 2026 TH2 - AZsPCs (Median Human Percentile): leaderboard
METR's Q4 2026 capability update (20 November 2026): where a model's best 48-hour run ranks among the human entrants of 15 past Al Zimmermann's Programming Contests (heuristic optimization problems that humans had about three months to work on), as the median percentile across the contests. 100 means beating every human entrant. Six models.
Metric: Median Percentile Among Human Participants. Source: metr.org. Saturation forecast: Around February 2027. 6 models tracked. Updated 21 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra | 98.45 | #10 |
| 2 | Claude Opus 5.5 | 97.49 | #9 |
| 3 | Claude Fable 5.1 | 96.44 | #20 |
| 4 | GPT-5.6 Sol | 89 | #37 |
| 5 | Claude Opus 4.5 | 61.86 | #160 |
| 6 | GPT-5 | 48 | #178 |
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-azspcs-median-human-percentile · How It Works · Data refreshed daily, snapshot 2026-11-21.