AA GDP.pdf - Criterion Pass Rate: leaderboard
Artificial Analysis's implementation of Surge AI's GDP.pdf, professional tasks over long real-world PDF documents graded on task-specific atomic criteria across ten domains. This board is the task-macro mean criterion pass rate, 0-100 (%), higher is better; the AA GDP.pdf board carries the stricter all-criteria pass rate.
Metric: Macro Criterion Pass Rate (%). Source: artificialanalysis.ai. Saturation forecast: Around February 2027. 242 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6.1 Sol (High) | 82.86 | #7 (GPT-6.1 Sol) |
| 2 | GPT-6 Astra (High) | 82.05 | #6 (GPT-6 Astra) |
| 3 | GPT-6 Astra (Medium) | 81.83 | #6 (GPT-6 Astra) |
| 4 | GPT-6 Astra (Max) | 81.81 | #6 (GPT-6 Astra) |
| 5 | GPT-6 Astra (xHigh) | 81.78 | #6 (GPT-6 Astra) |
| 6 | GPT-6.1 Sol (xHigh) | 81.42 | #7 (GPT-6.1 Sol) |
| 7 | GPT-6.1 Sol (Max) | 81.4 | #7 (GPT-6.1 Sol) |
| 8 | GPT-6 Astra (Low) | 80.81 | #6 (GPT-6 Astra) |
| 9 | Gemini 4 Carbon (High) | 80.77 | #1 (Gemini 4 Carbon) |
| 10 | GPT-6.1 Sol (Medium) | 80.57 | #7 (GPT-6.1 Sol) |
| 11 | Muse Wonder Full (Max) | 80.35 | #12 (Muse Wonder Full) |
| 12 | Kimi K3.1 Pro | 80.22 | #17 |
| 13 | Gemini 4 Argon (High) | 79.91 | #9 (Gemini 4 Argon) |
| 14 | GPT-6.1 Terra (Max) | 79.54 | #15 (GPT-6.1 Terra) |
| 15 | GPT-6.1 Sol (Low) | 79.27 | #7 (GPT-6.1 Sol) |
No result here: #3 Claude Opus 5.5, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=aa-gdp-pdf-criterion-pass-rate · How It Works · Data refreshed daily, snapshot 2026-11-05.