CLBench Life - Fragmented Information & Revisions: leaderboard
Metric: Solving Rate (%). Source: www.clbench.com. 29 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 (High) | 17.8 |
| 2 | GPT-5.4 (xHigh) | 17.4 |
| 3 | Claude Opus 4.6 (High) | 16.5 |
| 4 | Hy3-preview (High) | 15.9 |
| 5 | GPT-5.4 (High) | 15.8 |
| 6 | Gemini 3.1 Pro (Preview) (High) | 13.6 |
| 7 | Gemini 3.1 Pro (Preview) | 13.3 |
| 8 | DeepSeek V4 Pro (High) | 12.1 |
| 9 | Seed 2.0 Pro (High) | 12.1 |
| 10 | Kimi K2.5 | 11.9 |
| 11 | GPT-5.1 (High) | 11.9 |
| 12 | Kimi K2.5 (High) | 11.4 |
| 13 | Claude Opus 4.6 | 11.1 |
| 14 | GPT-5.4 | 10.4 |
| 15 | Seed 2.0 Pro | 10.4 |
Interactive version: theaggregate.ai/benchmark?slug=clbench-life-fragmented-information-revisions · How It Works · Data refreshed daily, snapshot 2026-09-19.