SuperCLUE-CoT - Overall: leaderboard
SuperCLUE-CoT evaluates chain-of-thought reasoning of Chinese LLMs, assessing step-by-step logical reasoning quality. Headline board of the newest workbook the superclueai.com vertical lists; score 0-100, higher is better.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Estimated already saturated. 34 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 | 77.66 | #190 |
| 2 | O1 Preview (2024-09-12) | 75 | #214 |
| 3 | QwQ 32B-Preview | 64.64 | #671 |
| 4 | DeepSeek R1 | 63.42 | #262 |
| 5 | Claude 3.5 Sonnet | 58.68 | #362 |
| 6 | Gemini 1.5 Pro (002) | 56.67 | #368 |
| 7 | GLM-4 Plus | 53.32 | #398 |
| 8 | Yi Lightning | 52.39 | #429 |
| 9 | GPT-4 Turbo | 50.79 | #417 |
| 10 | Qwen 2.5 72B Instruct | 47.9 | #469 |
| 11 | O1 Mini (2024-09-12) | 45.54 | #412 |
| 12 | Step 2 16K | 43.84 | #442 |
| 13 | ERNIE-4.0-Turbo-8K-Preview | 40.75 | #603 |
| 14 | Llama 3.1 405B Instruct | 40.7 | #470 |
| 15 | Llama 3.1 70B Instruct | 37.71 | #567 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-cot-overall · How It Works · Data refreshed daily, snapshot 2026-11-05.