MMStar
MMStar - comprehensive multimodal benchmark evaluating fine-grained visual reasoning.
Category
Vision & Language
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for MMStar
| Model | Score | Date | Verified |
|---|---|---|---|
| Qwen3.6-27B |
77.36%
|
22.04.2026 | Verified |
| Qwen3.5-397B-A17B |
100.00%
|
22.04.2026 | Unverified |
| Gemma4-31B-it |
38.68%
|
22.04.2026 | Unverified |
| Claude 4.5 Opus |
73.20
|
22.04.2026 | Unverified |
| Qwen3.6-35B-A3B |
70.75%
|
22.04.2026 | Unverified |
| Qwen3.5-27B |
73.58%
|
22.04.2026 | Unverified |
| Qwen3.5-4B |
48.11%
|
01.02.2026 | Verified |
| Qwen3.5-122B-A10B |
91.51%
|
01.02.2026 | Verified |
| Qwen3.5-35B-A3B |
82.08%
|
01.02.2026 | Unverified |
| Claude-Sonnet-4.5 |
5.66%
|
01.02.2026 | Unverified |
| GPT-5-mini 2025-08-07 |
8.49%
|
01.02.2026 | Unverified |
| Qwen3-235B-A22B |
51.89%
|
01.02.2026 | Unverified |
| Qwen3-VL-235B-A22B |
51.89%
|
01.02.2026 | Verified |