SuperCLUE-Ondevice - Video Understanding: leaderboard
SuperCLUE is the Chinese-language evaluation programme run by CLUE; its specialised boards each cover one professional domain and are graded by the organisers rather than self-reported. This board covers on-device models: video understanding.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Around October 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Seed 2.1 Pro | 85.54 | #67 |
| 2 | Gemini 3.6 Flash | 84.09 | #51 |
| 3 | Qwen 3.8 Max | 83.51 | #35 |
| 4 | Qwen 3.5 9B | 79.17 | #379 |
| 5 | Qwen 3.5 4B | 74.63 | #475 |
| 6 | GLM-5V Turbo | 67.22 | #166 |
| 7 | Gemma 4 E2B (IT) | 47.38 | #781 |
| 8 | gemma-4-E4B-it | 46 | #566 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-ondevice-video-understanding · How It Works · Data refreshed daily, snapshot 2026-10-20.