SuperCLUE General (March 2025) - Agent: leaderboard
SuperCLUE general leaderboard, March 2025 edition (2025年3月): agent (智能体Agent) score (总排行榜 sheet). SuperCLUE's periodic Chinese-language general evaluation uses a new question set each edition, so editions are separate boards. Score 0-100, higher is better.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Estimated already saturated. 36 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.5 (Preview) | 71.88 | #260 |
| 2 | DeepSeek R1 | 65.18 | #262 |
| 3 | Gemini 2.0 Pro (Preview 02-05) | 64.88 | #318 |
| 4 | DeepSeek V3 | 63.39 | #329 |
| 5 | Step 2 16K | 59.38 | #442 |
| 6 | O3 Mini (High) | 57.14 | #282 (O3 Mini) |
| 7 | Claude 3.7 Sonnet | 56.62 | #259 |
| 8 | Qwen 2.5 72B Instruct | 55.8 | #469 |
| 9 | Doubao-1.5-pro-32k-250115 | 54.46 | #320 |
| 10 | GLM-4 Plus | 49.55 | #398 |
| 11 | QwQ-32B | 48.66 | #422 |
| 12 | Llama 3.3 70B Instruct | 39.29 | #541 |
| 13 | DeepSeek R1 Distill Qwen 32B | 36.77 | #662 |
| 14 | Qwen 2.5 14B Instruct | 32.59 | #654 |
| 15 | GLM-4 9B Chat | 30.36 | #911 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-general-march-2025-agent · How It Works · Data refreshed daily, snapshot 2026-11-05.