Video Action Retrieval on SOVABench Intra-pair 1.0
53.9Pair-mAPGemini 2.5 Flash
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini 2.5 FlashCategory=API MLLM, Prompting Strategy=TASK-AWARE, Framework=MLLM-to-Embedding framework2026.01 | 53.9 | |
| MiniCPM-V 4.5Category=General MLLM, Prompting Strategy=TASK-AWARE, Framework=MLLM-to-Embedding framework2026.01 | 53.6 | |
| VideoLlama3 7BCategory=Video-MLLM, Prompting Strategy=TASK-AWARE, Framework=MLLM-to-Embedding framework2026.01 | 53.1 | |
| Gemini 2.5 FlashCategory=API MLLM, Prompting Strategy=GENERAL, Framework=MLLM-to-Embedding framework2026.01 | 52.6 | |
| Qwen3-VL 235B A22BCategory=API MLLM, Prompting Strategy=TASK-AWARE, Framework=MLLM-to-Embedding framework2026.01 | 52.6 | |
| MiniCPM-V 4.5Category=General MLLM, Prompting Strategy=GENERAL, Framework=MLLM-to-Embedding framework2026.01 | 52.5 | |
| InternVL3.5 8BCategory=General MLLM, Prompting Strategy=TASK-AWARE, Framework=MLLM-to-Embedding framework2026.01 | 52.4 | |
| VideoLlama3 7BCategory=Video-MLLM, Prompting Strategy=GENERAL, Framework=MLLM-to-Embedding framework2026.01 | 52.3 | |
| VideoChat-R1 7BCategory=Video-MLLM, Prompting Strategy=TASK-AWARE, Framework=MLLM-to-Embedding framework2026.01 | 51.9 | |
| Qwen3-VL 235B A22BCategory=API MLLM, Prompting Strategy=GENERAL, Framework=MLLM-to-Embedding framework2026.01 | 51.9 | |
| VideoChat-R1 7BCategory=Video-MLLM, Prompting Strategy=GENERAL, Framework=MLLM-to-Embedding framework2026.01 | 51.8 | |
| InternVL3.5 8BCategory=General MLLM, Prompting Strategy=GENERAL, Framework=MLLM-to-Embedding framework2026.01 | 51.7 | |
| VideoCLIPCategory=Contrastive Video-VLM2026.01 | 51.4 | |
| CLIP4ClipCategory=Contrastive Video-VLM2026.01 | 51.4 | |
| CLIP-ViT-L-14Category=Contrastive Image-VLM, Backbone=ViT-L-142026.01 | 51.3 | |
| MERUCategory=Contrastive Image-VLM2026.01 | 51.3 | |
| ActionCLIPCategory=Contrastive Video-VLM2026.01 | 51.3 | |
| SigLIP2-GiantCategory=Contrastive Image-VLM2026.01 | 51.1 | |
| VideoLLaVA 7BCategory=Video-MLLM, Prompting Strategy=GENERAL, Framework=MLLM-to-Embedding framework2026.01 | 51.1 | |
| VideoLLaVA 7BCategory=Video-MLLM, Prompting Strategy=TASK-AWARE, Framework=MLLM-to-Embedding framework2026.01 | 51 | |
| RandomCategory=Baseline2026.01 | 50.3 |