General Reasoning & Understanding on RealWorldQA (accuracy)
72.6Accuracy (RealWorldQA)V-STAR
Evaluation results
| Method | Links | |
|---|---|---|
| V-STARData Size=40k2026.04 | 72.6 | |
| LEADBase Model=VL-Rethinker-7B, Decoding Strategy=LEAD2026.03 | 71.2 | |
| ThinkLite-VLData Size=11k2026.04 | 70.1 | |
| VL-Rethinker-7BDecoding Strategy=Standard2026.03 | 69.3 | |
| VL-RethinkerData Size=39k2026.04 | 69.3 | |
| LEADBase Model=VL-Cogito-7B, Decoding Strategy=LEAD2026.03 | 69.2 | |
| Qwen2.5VL2026.04 | 68.8 | |
| VL-Cogito-7BDecoding Strategy=Standard2026.03 | 68.1 | |
| VL-CogitoData Size=80k2026.04 | 68.1 | |
| LEADBase Model=Vision-R1-7B, Decoding Strategy=LEAD2026.03 | 67.5 | |
| LEADBase Model=R1-Onevision-7B, Decoding Strategy=LEAD2026.03 | 66.4 | |
| LEADBase Model=OpenVLThinker-7B, Decoding Strategy=LEAD2026.03 | 65.3 | |
| SIDBase Model=R1-Onevision-7B, Decoding Strategy=SID2026.03 | 65.2 | |
| MemVRBase Model=R1-Onevision-7B, Decoding Strategy=MemVR2026.03 | 64.3 | |
| Vision-R1-7BDecoding Strategy=Standard2026.03 | 64.3 | |
| Vision-R1Data Size=210k2026.04 | 64.3 | |
| VCDBase Model=R1-Onevision-7B, Decoding Strategy=VCD2026.03 | 62.6 | |
| R1-Onevision-7BDecoding Strategy=Standard2026.03 | 62.5 | |
| OpenVLThinker-7BDecoding Strategy=Standard2026.03 | 62.3 | |
| OpenVLThinkerData Size=59.2k2026.04 | 62.3 | |
| R1-OnevisionData Size=155k2026.04 | 60.5 |