Image-Text Reasoning on MathVista
79.85AccuracyDAPO + ReMind
Evaluation Results
| Method | Links | |
|---|---|---|
| DAPO + ReMindBase Model=Qwen3-VL-8B-Instruct2026.06 | 79.85 | |
| GRPO + ReMindBase Model=Qwen3-VL-8B-Instruct2026.06 | 79.1 | |
| RLEPBase Model=Qwen3-VL-8B-Instruct2026.06 | 78.5 | |
| RePOBase Model=Qwen3-VL-8B-Instruct2026.06 | 78 | |
| DAPOBase Model=Qwen3-VL-8B-Instruct2026.06 | 77.6 | |
| ExGRPOBase Model=Qwen3-VL-8B-Instruct2026.06 | 77.3 | |
| GRPOBase Model=Qwen3-VL-8B-Instruct2026.06 | 76.2 | |
| Base ModelBase Model=Qwen3-VL-8B-Instruct2026.06 | 73.8 |