AA GDP.pdf: leaderboard
Artificial Analysis evaluates agent work involving PDF documents. Reports the percentage of tasks that satisfy all grading criteria, from the complete public model roster.
Metric: All Criteria Pass Rate (%). Source: artificialanalysis.ai. Saturation forecast: Around 2028. 242 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Wonder Full (Max) | 32.6 | #12 (Muse Wonder Full) |
| 2 | GPT-6 Astra (xHigh) | 32.2 | #6 (GPT-6 Astra) |
| 3 | GPT-6.1 Sol (High) | 32 | #7 (GPT-6.1 Sol) |
| 4 | GPT-6.1 Sol (xHigh) | 31.8 | #7 (GPT-6.1 Sol) |
| 5 | Kimi K3.1 Pro | 31.4 | #17 |
| 6 | GPT-6 Astra (Max) | 31 | #6 (GPT-6 Astra) |
| 7 | GPT-6.1 Sol (Max) | 31 | #7 (GPT-6.1 Sol) |
| 8 | GPT-6 Astra (High) | 31 | #6 (GPT-6 Astra) |
| 9 | Muse Wonder (Max) | 30.6 | #18 (Muse Wonder) |
| 10 | GPT-6 Astra (Medium) | 30.4 | #6 (GPT-6 Astra) |
| 11 | GPT-6 Astra (Low) | 30.4 | #6 (GPT-6 Astra) |
| 12 | GPT-6.1 Sol (Medium) | 30 | #7 (GPT-6.1 Sol) |
| 13 | GPT-6.1 Terra (Max) | 28.6 | #15 (GPT-6.1 Terra) |
| 14 | GPT-6.1 Terra (Medium) | 28.4 | #15 (GPT-6.1 Terra) |
| 15 | Gemini 4 Carbon (High) | 28.4 | #1 (Gemini 4 Carbon) |
No result here: #3 Claude Opus 5.5, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=aa-gdp-pdf · How It Works · Data refreshed daily, snapshot 2026-11-05.