Share your thoughts, 1 month free Claude Pro on usSee more

Proactive Assistance on Pro2Bench EgoProactive

68G-Mean F1

Qwen3.6-VL-27B

Updated 1mo ago

Evaluation Results

Method	Links
Qwen3.6-VL-27B 2026.06		68	44
Claude Opus 4.6 2026.06		57	43
Qwen3-VL-235B 2026.06		57	33
Llama 4-Maverick 2026.06		57	43
Llama 4-Maverick 2026.06		49	34
Llama 4-Maverick 2026.06		49	40
Gemini 3.1 Pro 2026.06		46	38
Qwen3.6-VL-27B 2026.06		28	24
Qwen3.6-VL-27B 2026.06		22	46
GPT-5.2 2026.06		0	-