Multi-modal Reasoning on CV-Bench
86.5Overall AccuracyRobix 7B Base
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Robix 7B Baseparameters=7B, type=Base2025.12 | 86.5 | — | |
| Lumo-1-Stage1training_stage=Stage 12025.12 | 86.36 | — | |
| RoboBrain-7B-2.0parameters=7B2025.12 | 85.81 | — | |
| Lumo-1-Stage2training_stage=Stage 22025.12 | 84.93 | — | |
| Qwen2.5-VL-32B-Instructparameters=32B, type=Instruct2025.12 | 81.78 | — | |
| Qwen2.5-VL-7B-Instructparameters=7B, type=Instruct2025.12 | 79.1 | — | |
| CORAModel=Qwen2.5-VL-7B2026.06 | 78.24 | 7.36 | |
| CORAModel=Qwen2-VL-7B2026.06 | 78.09 | 8.64 | |
| BaselineModel=Qwen2.5-VL-7B2026.06 | 77.75 | 8.08 | |
| BaselineModel=Qwen2-VL-7B2026.06 | 77.52 | 9.56 | |
| CORAModel=Qwen2.5-VL-3B2026.06 | 76.5 | 9.11 | |
| BaselineModel=Qwen2.5-VL-3B2026.06 | 76.04 | 10.43 | |
| CORAModel=Qwen2-VL-2B2026.06 | 68.39 | 12.44 | |
| BaselineModel=Qwen2-VL-2B2026.06 | 68.12 | 17.77 |