Multimodal Mathematical Reasoning on MathVista testmini
63ScoreInternVL-2-8B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| InternVL-2-8BBackbone=InternVL-2-8B2026.01 | 63 | 602.8 | — | |
| InternVL-2-26BBackbone=InternVL-2-26B2026.01 | 61.1 | 600.1 | — | |
| IXC-2.5-7BBackbone=IXC-2.5-7B2026.01 | 60.4 | 380.1 | — | |
| SLIBackbone=LLaVA-OV-7B2026.01 | 57.8 | 551.1 | — | |
| Qwen3-VL-8BSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 56.1 | — | 100 | |
| CLIBackbone=LLaVA-OV-7B2026.01 | 55.8 | 559.4 | — | |
| DeepStackBackbone=LLaVA-OV-7B2026.01 | 55.5 | 539.3 | — | |
| LLaVA-OV-7BBackbone=LLaVA-OV-7B2026.01 | 55.4 | 556.2 | — | |
| ResAdaptSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 52.5 | — | 42 | |
| ToMeSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 49.6 | — | 50 | |
| Qwen2.5-VL-7BSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 49.1 | — | 100 | |
| VisionZipSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 47.8 | — | 50 | |
| Random DropSetup=Zero-shot, Base Model=Qwen3-VL-8B2026.03 | 47.3 | — | 50 | |
| VisionZipSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 47.2 | — | 50 | |
| ResAdapt-RLSetup=Zero-shot, Base Model=Qwen2.5-VL-7B, RL fine-tuning=true2026.03 | 46.7 | — | 42 | |
| ToMeSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 46.2 | — | 50 | |
| ResAdaptSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 45.5 | — | 42 | |
| Random DropSetup=Zero-shot, Base Model=Qwen2.5-VL-7B2026.03 | 44.8 | — | 50 | |
| SLIBackbone=LLaVA-OV-0.5B2026.01 | 37.9 | 393.7 | — | |
| CLIBackbone=LLaVA-OV-0.5B2026.01 | 37.1 | 395.6 | — | |
| LLaVA-OV-0.5BBackbone=LLaVA-OV-0.5B2026.01 | 36.3 | 391 | — | |
| VILA-13BBackbone=VILA-13B2026.01 | 33.7 | 396.2 | — | |
| DeepStackBackbone=LLaVA-OV-0.5B2026.01 | 32.5 | 380.9 | — |