List your company

Long Video Understanding and Reasoning on VideoMME w/ subtitle

81.3AccuracyGemini 1.5 Pro
9.74828.32446.965.476Nov 25, 2025

Evaluation results

MethodLinks
Gemini 1.5 ProReasoning Prompt=false, Tool Calling=false2025.1181.355.2
GPT-4oReasoning Prompt=false, Tool Calling=false2025.1177.251.5
LongVT-7B-RFTReasoning Prompt=true, Tool Calling=true, Frame Sampling Regime=Dense2025.116747.7
LongVT-7B-RLReasoning Prompt=true, Tool Calling=true, Frame Sampling Regime=Sparse2025.1166.143.7
LongVT-7B-RLReasoning Prompt=true, Tool Calling=true, Frame Sampling Regime=Dense2025.1166.146.6
LongVT-7B-SFTReasoning Prompt=true, Tool Calling=true, Frame Sampling Regime=Dense2025.1164.944.1
Qwen2.5-VL-7BReasoning Prompt=false, Tool Calling=false, Frame Sampling Regime=Dense2025.1164.346
Qwen2.5-VL-7BReasoning Prompt=false, Tool Calling=false, Frame Sampling Regime=Sparse2025.1162.637.2
Video-R1-7BReasoning Prompt=true, Tool Calling=false, Frame Sampling Regime=Sparse2025.116142.6
Video-Thinker-7BReasoning Prompt=true, Tool Calling=false, Frame Sampling Regime=Sparse2025.116142.6
VideoRFT-7BReasoning Prompt=true, Tool Calling=false, Frame Sampling Regime=Sparse2025.1160.942.3
Video-Thinker-7BReasoning Prompt=true, Tool Calling=false, Frame Sampling Regime=Dense2025.1160.842.9
Video-R1-7BReasoning Prompt=true, Tool Calling=false, Frame Sampling Regime=Dense2025.1160.542.7
VideoRFT-7BReasoning Prompt=true, Tool Calling=false, Frame Sampling Regime=Dense2025.1149.237
LongVT-7B-SFTReasoning Prompt=true, Tool Calling=true, Frame Sampling Regime=Sparse2025.1112.536.2