SuperCLUE-VLM (December 2025) - Overall: leaderboard
SuperCLUE-VLM evaluates vision-language models in Chinese on basic visual cognition, visual reasoning and applied visual tasks. December 2025 edition (2025年12月): headline 总分 of the 总榜成绩 sheet. Each dated workbook is a separate evaluation round, kept apart from the current board. Score 0-100, higher is better.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Around February 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 83.64 | #82 |
| 2 | Qwen 3 VL 235B A22B (Thinking) | 71.95 | #236 |
| 3 | Claude Opus 4.5 (20251101) | 71.44 | #81 |
| 4 | GPT-5.2 (High) | 69.16 | #115 (GPT-5.2) |
| 5 | GLM-4.6V | 67.68 | #330 |
| 6 | Step-3 | 62.94 | #380 |
| 7 | Qwen 3 VL 8B Instruct | 61.64 | #419 |
| 8 | Grok 4.1 Fast (Non-reasoning) | 45.57 | #227 (Grok 4.1 Fast) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-vlm-december-2025-overall · How It Works · Data refreshed daily, snapshot 2026-11-05.