ASI-Bench (Claude Code) - B2 Method Hints: leaderboard
Metric: Scientific Score (0-100). Source: huggingface.co. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 5 (Max) | 45.8 |
| 2 | Gemini 3.8 Flash | 39.13 |
| 3 | Claude Opus 4.8 (Max) | 36.25 |
| 4 | GLM-5.3 (Max) | 35.45 |
| 5 | Kimi K3 (Max) | 35.24 |
| 6 | Qwen 3.8 Max (Max) | 32.51 |
| 7 | GLM-5.2 (Max) | 30.81 |
| 8 | GLM-5.3 Flash (Max) | 28.66 |
| 9 | DeepSeek V4 Flash (Max) | 26.49 |
| 10 | DeepSeek V4 Pro (Max) | 18.79 |
Interactive version: theaggregate.ai/benchmark?slug=asi-bench-claude-code-b2-method-hints · How It Works · Data refreshed daily, snapshot 2026-09-19.