Updated daily at 05:00 UTC
Benchmark
CL-bench
Measures: Context learning: 1,899 expert tasks where the model must learn knowledge absent from pretraining (invented legal systems, game rules, financial instruments) from its context and apply it; a task counts only if every rubric passes.
- Maintainer
- Tencent Hunyuan and Fudan University NLP group
- Introduced
- 2026
- Status
- Active
- Saturation
- Headroom
- Typical provenance
- Independent
- In scoring set
- No
Leaderboard
Every result we hold, whoever ran it. Only independent and third-party rows count towards a score.
| Model | Tasks solved | Provenance | Source | Evaluated |
|---|---|---|---|---|
| MiniMax-M2.5 | 11.4% | Source not stated | Epoch AI via Epoch AI | — |
| Kimi K2 Thinking | 11.9% | Source not stated | Epoch AI via Epoch AI | — |
| DeepSeek-V3.2 | 12.4% | Source not stated | Epoch AI via Epoch AI | — |
| DeepSeek-V3.2-Exp DeepSeek · Thinking effort | 13.2% | Source not stated | Epoch AI via Epoch AI | — |
| Qwen3-Max | 14.5% | Source not stated | Epoch AI via Epoch AI | — |
| MiMo-V2-Pro | 15.7% | Source not stated | Epoch AI via Epoch AI | — |
| Gemini 3 Pro | 15.8% | Source not stated | Epoch AI via Epoch AI | — |
| GLM-4.7 | 15.9% | Source not stated | Epoch AI via Epoch AI | — |
| Kimi K2 Thinking | 17.6% | Source not stated | Epoch AI via Epoch AI | — |
| o3 OpenAI · High effort | 17.8% | Source not stated | Epoch AI via Epoch AI | — |
| GPT-5.2 OpenAI · High effort | 18.1% | Source not stated | Epoch AI via Epoch AI | — |
| GPT-5.2 OpenAI · Unknown effort | 18.2% | Source not stated | Epoch AI via Epoch AI | — |
| GLM-5 | 18.7% | Source not stated | Epoch AI via Epoch AI | — |
| Kimi K2.5 | 19.3% | Source not stated | Epoch AI via Epoch AI | — |
| Qwen3.5-Plus | 19.8% | Source not stated | Epoch AI via Epoch AI | — |
| Qwen3.6-Plus | 20.3% | Source not stated | Epoch AI via Epoch AI | — |
| Claude Opus 4.6 Anthropic · Unknown effort | 20.7% | Source not stated | Epoch AI via Epoch AI | — |
| Gemini 3.1 Pro | 20.8% | Source not stated | Epoch AI via Epoch AI | — |
| Claude Opus 4.5 Anthropic · Unknown effort | 21.1% | Source not stated | Epoch AI via Epoch AI | — |
| GPT-5.1 OpenAI · Unknown effort | 21.1% | Source not stated | Epoch AI via Epoch AI | — |
| Grok 4.20 | 22.2% | Source not stated | Epoch AI via Epoch AI | — |
| GPT-5.1 OpenAI · High effort | 23.7% | Source not stated | Epoch AI via Epoch AI | — |
| GPT-5.4 OpenAI · Xhigh effort | 27.9% | Source not stated | Epoch AI via Epoch AI | — |
- Epoch AI, 'Capabilities & benchmarking', epoch.ai/benchmarks, CC BY 4.0.