General MCQA on SeedBench
71.5ScoreScaled base VLM 7B + MERIT (Proposed, 2D)
Evaluation results
| Method | Links | |
|---|---|---|
| Scaled base VLM 7B + MERIT (Proposed, 2D)Train Data=3.6M + 1.6M2026.06 | 71.5 | |
| Scaled base VLM 7B + Joint FFTTrain Data=3.6M + 1.6M2026.06 | 71.3 | |
| Base VLM 7B + MERIT (Proposed, 2D)Train Data=0.7M + 1.6M2026.06 | 70.6 | |
| Base VLM 7B + Joint FFTTrain Data=0.7M + 1.6M2026.06 | 70.1 | |
| Scaled base VLM 7BTrain Data=3.6M2026.06 | 69.8 | |
| LLaVA-1.5-13BTrain Data=0.7M2026.06 | 68.2 | |
| LLaVA-1.5-7BTrain Data=0.7M2026.06 | 65.9 | |
| Base VLM 7BTrain Data=0.7M2026.06 | 64.5 | |
| LLaVA-7BTrain Data=0.6M2026.06 | 37 |