LLM Stats (MMBench-V1.1): leaderboard
Systematic multimodal evaluation with 2,974 multiple-choice questions spanning 20 ability dimensions like object localization, attribute reasoning, and spatial understanding.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 122B A10B | 92.8 | #173 |
| 2 | Qwen 3.6 35B A3B | 92.8 | #226 |
| 3 | Qwen 3.5 27B | 92.6 | #187 |
| 4 | Qwen 3.6 27B | 92.3 | #170 |
| 5 | Qwen 3.5 35B A3B | 91.5 | #256 |
| 6 | Qwen 3 VL 32B (Thinking) | 90.8 | #294 (Qwen 3 VL 32B) |
| 7 | Qwen 3 VL 235B A22B (Thinking) | 90.6 | #238 (Qwen 3 VL 235B A22B) |
| 8 | Qwen 3 VL 235B A22B Instruct | 89.9 | #273 |
| 9 | Qwen 3 VL 30B A3B (Thinking) | 88.9 | #351 (Qwen 3 VL 30B A3B) |
| 10 | Qwen 3 VL 8B (Thinking) | 87.5 | |
| 11 | Qwen 3 VL 30B A3B Instruct | 87 | #384 |
| 12 | Qwen 3 VL 4B (Thinking) | 86.7 | #480 (Qwen 3 VL 4B) |
| 13 | Qwen 3 VL 4B Instruct | 85.1 | #515 |
| 14 | Qwen 3 VL 8B Instruct | 85 | #420 |
| 15 | Qwen2.5-Omni-7B | 81.8 | #624 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-mmbench-v1-1 · How It Works · Data refreshed daily, snapshot 2026-10-20.