Audio-Visual Question Answering on Video-Holmes (Score)
59.9ScoreGemini 2.5 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 2.5 ProModel Category=Proprietary Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 59.9 | |
| Gemini 2.5 FlashModel Category=Proprietary Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 52.8 | |
| Gemini 2.0 FlashModel Category=Proprietary Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 51.6 | |
| Omni-Captioner-7BModel Category=Open-Source Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 48.8 | |
| video-SALMONN 2-7BModel Category=Open-Source Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 42.9 | |
| Qwen2.5-Omni-7BModel Category=Open-Source Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 35.7 | |
| VideoLLaMA 2-7BModel Category=Open-Source Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 33.5 | |
| video-SALMONN-13BModel Category=Open-Source Models, QA Backbone=GPT-4o, Evaluation Approach=caption-to-QA cascade2025.10 | 31.4 |