| Benchmark | Model | Rank | Results |
|---|---|---|---|
| video-question-answering-on-tomato | Molmo2 (SigLIP2 So400M + Qwen3-4B) baseline | #6 | Accuracy: 36.5 |
| video-question-answering-on-tempcompass-overall | Molmo2 (SigLIP2 So400M + Qwen3-4B) baseline | #11 | Accuracy: 69.9 |
| video-question-answering-on-mvbench | Molmo2 (SigLIP2 So400M + Qwen3-4B) with STTS (30%) | #19 | Avg.: 72.7 |
| video-question-answering-on-mlvu | Molmo2 (SigLIP2 So400M + Qwen3-4B) with STTS (30%) + Test-Time Scaling (92 frames) | #34 | Accuracy: 70.5 |