MATH-Vision
9 models
Top 5 Models Performance
| Qwen: qwen3.8-27b | ######################################## | 90 |
| Google: Gemma 4 31B Instruct | ###################################### | 85.6 |
| Google: Gemma 4 12B Instruct | ################################### | 79.7 |
| NVIDIA Corporation: nvidia-nemotron-3.5-super-vl-120b-a12b | ################################### | 78.16 |
| Google: diffusiongemma-26b-a4b-it | ############################### | 70.5 |
| Rank | Model | Score | |
|---|---|---|---|
| 🇨🇳 | 1 | Qwen: qwen3.8-27b通义千问 | 90 |
| 🇺🇸 | 2 | Google: Gemma 4 31B Instruct | 85.6 |
| 🇺🇸 | 3 | Google: Gemma 4 12B Instruct | 79.7 |
| 🇺🇸 | 4 | NVIDIA Corporation: nvidia-nemotron-3.5-super-vl-120b-a12b | 78.16 |
| 🇺🇸 | 5 | Google: diffusiongemma-26b-a4b-it | 70.5 |
| 🇨🇳 | 6 | Intern Language Models: intern-s1 | 62.5 |
| 🇺🇸 | 7 | Google: Gemma 4 E4B Instruct | 59.5 |
| 🇨🇳 | 8 | Qwen: qwen2.5-vl-3b-instruct通义千问 | 22.86 |
| ❓ | 9 | tfjiangxilab: tianjiangzhuge_3b | 22.14 |