CL-bench
Rank #29 of 51 · 19 models scored · top score 27.9
unknown
Strongest on contamination resistance, weakest on discrimination. Caveat: the public test set leaves contamination on the table.
Weighted composite. Each signal is computed from the source-backed leaderboard below.
Does it still separate models?
How far from ceiling / clustered at the top?
Public vs held-out; training-leak risk.
Apples-to-apples, or mixed / vendor-optimized?
How old is the benchmark?
| # | Model | Score | Evidence | Measured |
|---|---|---|---|---|
| 1 | GPT-5.4 | 27.9 | unverified | 2026-03-05 |
| 2 | GPT-5.1 | 23.7 | unverified | 2025-11-13 |
| 3 | Grok 4.20 | 22.2 | unverified | 2026-02-17 |
| 4 | Claude Opus 4.5 | 21.1 | unverified | 2025-11-24 |
| 5 | Gemini 3.1 Pro | 20.8 | unverified | 2026-02-19 |
| 6 | Claude Opus 4.6 | 20.7 | unverified | 2026-02-05 |
| 7 | Qwen 3.6 Plus | 20.3 | unverified | 2026-04-01 |
| 8 | Qwen 3.5 Plus (hosted 397B-A17B) | 19.8 | unverified | 2026-02-16 |
| 9 | Kimi K2.5 | 19.3 | unverified | 2026-02-02 |
| 10 | GLM-5 | 18.7 | unverified | 2026-02-11 |
| 11 | GPT-5.2 | 18.2 | unverified | 2025-12-11 |
| 12 | o3 | 17.8 | unverified | 2024-12-20 |
| 13 | Kimi K2 Thinking | 17.6 | unverified | 2025-11-06 |
| 14 | GLM-4.7 | 15.9 | unverified | 2025-12-22 |
| 15 | Gemini 3 Pro | 15.8 | unverified | 2025-11-18 |
| 16 | Qwen3-Max | 14.5 | unverified | 2025-09-05 |
| 17 | DeepSeek-V3.2-Exp | 13.2 | unverified | 2025-09-29 |
| 18 | DeepSeek-V3.2 | 12.4 | unverified | 2025-12-01 |
| 19 | MiniMax-M2.5 | 11.4 | unverified | 2026-02-12 |