CLASHarness v8 - Pass Rate (Hard): leaderboard
Metric: Episode pass rate, hard episodes (%). Source: github.com. Saturation forecast: Around 2029. 12 models tracked. Updated 18 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-6 Astra | High | 18.6 | #7 |
| 2 | GPT-5.6 Sol | High | 13.6 | #33 |
| 3 | Claude Opus 5 | High | 13.6 | #24 |
| 4 | Claude Opus 4.8 | Medium | 12.9 | #62 |
| 5 | Kimi K3 | High | 12.1 | #60 |
| 6 | GPT-5.6 Luna | Low | 12.1 | #129 |
| 7 | GPT-5.6 Terra | Medium | 10 | #69 |
| 8 | Gemini 3.7 Flash | Low | 9.3 | #38 |
| 9 | Gemini 3.6 Flash | Low | 9.3 | #77 |
| 10 | Claude Sonnet 5 | High | 7.9 | #116 |
| 11 | Gemini 3.1 Pro (Preview) | Medium | 6.4 | #99 |
| 12 | Gemini 3.5 Flash | Low | 5.7 | #106 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=clasharness-v8-pass-rate-hard · How It Works · Data refreshed daily, snapshot 2026-11-18.