List your company

Video Question Answering on TemporalBench

49.8VYRBaseline (Qwen3-VL-8B-Instruct)
-1.57611.76225.138.438Apr 19, 2026

Evaluation results

MethodLinks
Baseline (Qwen3-VL-8B-Instruct)Backbone=Qwen3-VL-8B-Instruct2026.0449.887.988.84.26112.11.683.280.71.3232.0411.1630.8171.86824.2
Baseline+CoTBackbone=Qwen3-VL-8B-Instruct, Chain-of-Thought=true2026.0448.187.488.34.26612.62.283.178.41.3350.811.2130.6540.29932.7
Baseline+SFTBackbone=Qwen3-VL-8B-Instruct, Supervised Fine-Tuning=true2026.044483.884.74.26816.21.680.173.81.2831.0141.6721.130.45534.8
Qwen3-VL-30BParameters=30B2026.0443.680.180.74.27515.72.782.372.81.2710.3651.5641.8941.36738.3
Baseline+CoT+SFTBackbone=Qwen3-VL-8B-Instruct, Chain-of-Thought=true, Supervised Fine-Tuning=true2026.0439.976.377.74.29723.713.374.466.61.1041.2171.5361.8130.51639.6
Internvl3.5-VL-241BParameters=241B2026.0438.476.779.14.26125.91.684.173.61.1430.6711.6132.0160.39840.5
Qwen3-VL-235BParameters=235B2026.0436.673.275.84.25726.71.282.671.71.0790.5911.6571.9460.21441.6
Gemini-3.1-Pro2026.0415.23838.24.24261.99.645.820.50.3310.3740.3641.1350.21662.5
VTHM-MoE2026.040.447.347.34.22352.79.649.32.20.0760.3070.7570.3670.89253.1