MM-MT-Bench
A multi-turn LLM-as-a-judge evaluation benchmark for testing multimodal instruction-tuned models' ability to follow user instructions in multi-turn dialogues and answer open-ended questions in a zero-shot manner.
17rows
scoreprimary metric
2026-05-06sampled
Metadata
Metrics
Score, Normalized Score
| Rank | Subject | Score | Model Match | Provenance | Sampled |
|---|---|---|---|---|---|
| 1 | Mistral Large 3 | 84.90 | — | Self-reported | 2026-05-06 |
| 2 | Qwen3 VL 235B A22B Instruct | 8.50 |  Qwen3 VL 235B A22B Instruct qwen-qwen3-vl-235b-a22b-instruct | Self-reported | 2026-05-06 |
| 2 | Qwen3 VL 235B A22B Thinking | 8.50 | Qwen3 VL 235B A22B Thinking qwen-qwen3-vl-235b-a22b-thinking | Self-reported | 2026-05-06 |
| 4 | Qwen3 VL 32B Instruct | 8.40 | Qwen3 VL 32B Instruct qwen-qwen3-vl-32b-instruct | Self-reported | 2026-05-06 |
| 5 | Qwen3 VL 32B Thinking | 8.30 | — | Self-reported | 2026-05-06 |
| 6 | Qwen3 VL 30B A3B Instruct | 8.10 | Qwen3 VL 30B A3B Instruct qwen-qwen3-vl-30b-a3b-instruct | Self-reported | 2026-05-06 |
| 7 | Qwen3 VL 8B Thinking | 8 | Qwen3 VL 8B Thinking qwen-qwen3-vl-8b-thinking | Self-reported | 2026-05-06 |
| 8 | Qwen3 VL 30B A3B Thinking | 7.90 | Qwen3 VL 30B A3B Thinking qwen-qwen3-vl-30b-a3b-thinking | Self-reported | 2026-05-06 |
| 9 | Qwen3 VL 8B Instruct | 7.70 | Qwen3 VL 8B Instruct qwen-qwen3-vl-8b-instruct | Self-reported | 2026-05-06 |
| 9 | Qwen3 VL 4B Thinking | 7.70 | — | Self-reported | 2026-05-06 |
| 11 | Qwen3 VL 4B Instruct | 7.50 | — | Self-reported | 2026-05-06 |
| 12 | Pixtral Large | 0.74 | — | Self-reported | 2026-05-06 |
| 13 | Pixtral-12B | 0.60 | — | Self-reported | 2026-05-06 |
| 14 | MiniStral 3 (14B Instruct 2512) | 0.08 | — | Self-reported | 2026-05-06 |
| 15 | Ministral 3 (8B Instruct 2512) | 0.08 | — | Self-reported | 2026-05-06 |
| 16 | Ministral 3 (3B Instruct 2512) | 0.08 | — | Self-reported | 2026-05-06 |
| 17 | Qwen2.5-Omni-7B | 0.06 | — | Self-reported | 2026-05-06 |
No matching rows.