LLM Stats (MMBench): leaderboard
Multimodal Benchmark with 2,974 multiple-choice questions spanning 20 ability dimensions from coarse perception to fine-grained reasoning.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Step3 VL 10B | 91.8 | #485 |
| 2 | Qwen 2.5 VL 72B Instruct | 88 | #377 |
| 3 | Phi-4 Multimodal Instruct | 86.7 | #906 |
| 4 | Qwen 2.5 VL 7B Instruct | 84.3 | #654 |
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 · How It Works · Data refreshed daily, snapshot 2026-10-20.