Video Question Answering on VideoMME (Aspect Metrics) benchmark leaderboard · SOTA2 Research

Video Question Answering on VideoMME (Aspect Metrics)

30.9VYRBaseline+CoT+SFT

Evaluation Results

Method
Baseline+CoT+SFT2026.04	30.9	—	—	—	—	40.7	42.2	3.892	55.9	15.8	60.4	35.4	0.466	1.162	1.313	1.753	1.307	53.9
Baseline+SFT2026.04	20.7	—	—	—	—	60.2	62.3	3.573	29.7	20.2	64.8	37.6	1.322	0.492	1.454	1.747	1.441	51.6
Baseline (Qwen3-VL-8B-Instruct)2026.04	19.3	—	—	—	—	61.7	63.7	3.571	28.4	21.1	65.6	39	1.376	0.483	1.709	1.447	0.837	31
Baseline+CoT2026.04	19	—	—	—	—	56	57.4	3.543	29	14.6	61.1	38.4	1.272	1.066	1.971	1.601	0.212	44.1
Internvl3.5-VL-241B2026.04	14.5	—	—	—	—	57.3	60.5	3.356	38.4	39.7	64.5	25.3	1.189	0.469	1.919	1.386	1.469	51.5
Qwen3-VL-30B2026.04	13.8	—	—	—	—	58.7	62.1	3.707	34.7	38.8	63.8	27.1	1.211	0.628	1.924	1.54	1.123	50.4
Qwen3-VL-235B2026.04	12.7	—	—	—	—	55.6	57.3	3.274	42.6	40.7	60.4	26.1	1.147	0.341	1.832	1.476	-1.316	53.7
VTHM-MoE2026.04	4.1	—	—	—	—	39.8	41.7	3.618	42.6	25.6	50.2	8.2	0.31	0.885	0.362	1.815	0.773	61.4
Gemini-3.1-Pro2026.04	0.7	—	—	—	—	13.5	13.2	3.387	60.1	11.4	28.7	1.6	0.045	0.388	0.698	1.336	0.212	59.8
LLoVi2025.01	—	52.2	55.3	51.8	50.8	—	—	—	—	—	—	—	—	—	—	—	—	—
video-grounded entailment tree reasoning2025.01	—	55.9	53.8	54	50.8	—	—	—	—	—	—	—	—	—	—	—	—	—
VideoTree2025.01	—	55.7	54.3	54.2	52.6	—	—	—	—	—	—	—	—	—	—	—	—	—