HANDBOOK.md Agents: leaderboard
Metric: Score (%). Source: surgehq.ai. 58 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5 (Adaptive Reasoning, Max Effort) | 36.2 |
| 2 | Claude Fable 5 | 34.2 |
| 3 | Claude Opus 5 (Adaptive Reasoning, Max Effort) | 32.3 |
| 4 | Claude Opus 5 | 29.6 |
| 5 | Grok 4.6 (xHigh) | 27.3 |
| 6 | DeepSeek V4 Pro (Max) | 26.9 |
| 7 | Grok 4.6 | 25.8 |
| 8 | GPT-5.6 Sol (Max) | 23.5 |
| 9 | Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | 21.9 |
| 10 | GPT-5.5 | 21.5 |
| 11 | GPT-5.6 Sol | 21.5 |
| 12 | GPT-5.5 (xHigh) | 21.5 |
| 13 | DeepSeek V4 Pro | 19.6 |
| 14 | Claude Opus 4.8 | 18.9 |
| 15 | Qwen 3.8 Max | 16.5 |
Interactive version: theaggregate.ai/benchmark?slug=handbook-md-agents · How It Works · Data refreshed daily, snapshot 2026-09-05.