List your company

Skill Evaluation on ALFRED

78.01Object Perception (Grounding) AccuracyBC (InternVL3-8B)
19.957235.028650.165.1714Apr 21, 2026

Evaluation results

MethodLinks
BC (InternVL3-8B)Backbone=InternVL3-8B2026.0478.0173.9471.88363.259.276398.53
Ours (InternVL3-8B)Backbone=InternVL3-8B2026.0475.8481.7399.498.696.889.0989.498.2
Gemini-2.5-ProModel Class=Closed2026.0463.5360.7586.280.285.4331.0881.985.76
GPT-o3Model Class=Closed2026.0457.2829.5594.880.686.7531.7679.677.41
GPT-5Model Class=Closed2026.0451.4524.3492.273.881.4828.8680.171.52
Qwen3-VL-8BModel Class=Open2026.0448.9951.7989.480.276.567.1881.545.34
Gemini-2.5-FlashModel Class=Closed2026.044543.1588.666.869.4532.5379.974.8
Claude-4.5-SonnetModel Class=Closed2026.0438.051.5490.865.851.662.5581.246.15
InternVL3-8BModel Class=Open2026.0433.726.6871.877.468.848.2674.24.42
InternVL2.5-8BModel Class=Open2026.0430.98.796776.268.747.3960.80.82
Qwen2.5-VL-7BModel Class=Open2026.0428.151.82847348.7849.6176.439.44
LLaVA-OneVision-7BModel Class=Open2026.0422.1918.1969.66871.0544.8869.56.87