Multi-modal Understanding on MMBench V1.1
87.03AccuracyInternVL3.5-38B
Evaluation Results
| Method | Links | |
|---|---|---|
| InternVL3.5-38BModel Type=Understanding-only MLLM2025.12 | 87.03 | |
| Qwen3VL-32BModel Type=Understanding-only MLLM2025.12 | 86.11 | |
| GPT-5Model Type=Understanding-only MLLM2025.12 | 84.25 | |
| Qwen3VL-8BModel Type=Understanding-only MLLM2025.12 | 84.25 | |
| InternVL3.5-8BModel Type=Understanding-only MLLM2025.12 | 83.33 | |
| GPT-4oModel Type=Understanding-only MLLM2025.12 | 82.4 | |
| LLaVANextData=OmniAlign-Vmix, LLM=Qwen2.5-32B2025.02 | 80.6 | |
| COOPERModel Type=Unified MLLM2025.12 | 80.55 | |
| LLaVANextData=LLaVANext-778k, LLM=Qwen2.5-32B2025.02 | 79.3 | |
| Gemma3-27BModel Size=27B2025.12 | 78.3 | |
| BAGEL-REModel Type=Unified MLLM, Variant=Reasoning Enhancement2025.12 | 77.77 | |
| LLaVANextData=LLaVANext-778k, LLM=InternLM2.5-7B2025.02 | 75.1 | |
| Gemma3-4B + AuditDMModel Size=4B, Auditing Component=AuditDM2025.12 | 75 | |
| LLaVANextData=OmniAlign-Vmix, LLM=InternLM2.5-7B2025.02 | 74.9 | |
| Gemma3-12BModel Size=12B2025.12 | 73.8 | |
| LLaVAData=OmniAlign-Vmix, LLM=InternLM2.5-7B2025.02 | 73.7 | |
| LLaVAData=LLaVANext-778k, LLM=InternLM2.5-7B2025.02 | 73.6 | |
| BAGELModel Type=Unified MLLM2025.12 | 72.22 | |
| BAGEL-PEModel Type=Unified MLLM, Variant=Perception Enhancement2025.12 | 69.44 | |
| Gemma3-4BModel Size=4B2025.12 | 67.6 | |
| Janus-Pro-7BModel Type=Unified MLLM2025.12 | 60.18 | |
| Liquid-7BModel Type=Unified MLLM2025.12 | 41.66 |