Benchmark
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…
- Modality
- multimodal
- Categories
- multimodal, communication
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 17
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| MiniStral 3 (14B Instruct 2512) | Mistral | 0.0849 | 2025-12-04 | evidence |
| Ministral 3 (3B Instruct 2512) | Mistral | 0.0783 | 2025-12-04 | evidence |
| Ministral 3 (8B Instruct 2512) | Mistral | 0.0808 | 2025-12-04 | evidence |
| Mistral Large 3 | Mistral | 84.9 | 2025-09-01 | evidence |
| Pixtral Large | Mistral | 0.74 | 2024-11-18 | evidence |
| Pixtral-12B | Mistral | 0.605 | 2024-09-17 | evidence |
| Qwen2.5-Omni-7B | Qwen | 0.06 | 2025-03-27 | evidence |
| Qwen3 VL 235B A22B Instruct | Qwen | 8.5 | 2025-09-22 | evidence |
| Qwen3 VL 235B A22B Thinking | Qwen | 8.5 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Instruct | Qwen | 8.1 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Thinking | Qwen | 7.9 | 2025-09-22 | evidence |
| Qwen3 VL 32B Instruct | Qwen | 8.4 | 2025-09-22 | evidence |
| Qwen3 VL 32B Thinking | Qwen | 8.3 | 2025-09-22 | evidence |
| Qwen3 VL 4B Instruct | Qwen | 7.5 | 2025-09-22 | evidence |
| Qwen3 VL 4B Thinking | Qwen | 7.7 | 2025-09-22 | evidence |
| Qwen3 VL 8B Instruct | Qwen | 7.7 | 2025-09-22 | evidence |
| Qwen3 VL 8B Thinking | Qwen | 8.0 | 2025-09-22 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.