Video Question Answering on TempCompass
76.9Multi-Choice AccuracyV-JEPA 2
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| V-JEPA 2Params Enc / LLM=1B/8B, Evaluation Protocol=zero-shot2025.06 | 76.9 | — | — | — | |
| Keye-VL-1.5Model Class=8B2026.05 | 75.5 | — | — | — | |
| LLaVA-OV-2Model Class=8B2026.05 | 74.5 | — | — | — | |
| Qwen3-VLModel Class=8B2026.05 | 74.3 | — | — | — | |
| STORMFrames=322026.05 | 74.3 | — | — | — | |
| GPT-4oFrames=322026.05 | 73.75 | — | — | — | |
| Qwen2.5-VL-7B-Instruct(CoT)Frames=322026.05 | 72.8 | — | — | — | |
| PLM 8BParams Enc / LLM=1B/8B, Evaluation Protocol=zero-shot2025.06 | 72.7 | — | — | — | |
| PLMModel Class=8B2026.05 | 72.7 | — | — | — | |
| Video-R1-7BFrames=322026.05 | 72.6 | — | — | — | |
| Qwen2.5VLParams Enc / LLM=1B/7B, Evaluation Protocol=zero-shot2025.06 | 71.7 | — | — | — | |
| InternVL-3.5Model Class=8B2026.05 | 70.4 | — | — | — | |
| Qwen2.5-VL-7B-SFTFrames=322026.05 | 69.9 | — | — | — | |
| Video-R1-7B-wo-imageFrames=322026.05 | 69.8 | — | — | — | |
| InternVL-2.5Params Enc / LLM=300M/7B, Evaluation Protocol=zero-shot2025.06 | 68.3 | — | — | — | |
| LongVT-RFTFrames=32, method_variant=think with tools2026.05 | 68.3 | — | — | — | |
| Qwen2VLParams Enc / LLM=675M/7B, Evaluation Protocol=zero-shot2025.06 | 67.9 | — | — | — | |
| Video-Thinker-7BFrames=32, method_variant=think with video2026.05 | 67.8 | — | — | — | |
| Gemini-1.5-ProFrames=322026.05 | 67.1 | — | — | — | |
| LLaVA-Next-VideoFrames=322026.05 | 66.74 | — | — | — | |
| KangarooFrames=322026.05 | 62.5 | — | — | — | |
| LLaVA-OV-1.5Model Class=8B2026.05 | 57.5 | — | — | — | |
| LongVAFrames=322026.05 | 56.9 | — | — | — | |
| VideoLLaVAFrames=322026.05 | 49.77 | — | — | — | |
| VideoChat2Frames=322026.05 | 48.8 | — | — | — | |
| LLaMA-VIDFrames=322026.05 | 45.6 | — | — | — | |
| TrajViTLLM=Llama-3, Setup=Linear layer connector, Pretraining Data=8M video + 50M image2025.05 | 40 | 50.6 | 59.9 | 31.1 | |
| ViT3DLLM=Llama-3, Setup=Linear layer connector, Pretraining Data=8M video + 50M image2025.05 | 39.1 | 50.1 | 58.4 | 30.4 |