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

Desktop Cleaning on Real-world robot tasks 1.0 (test)

8Success Rate

GPT-4o

Updated 7d ago

Evaluation Results

Method	Links
GPT-4o 2026.05		8
GPT-5 2026.05		7
Claude-3.7 2026.05		6
Claude-4.6 2026.05		6
Qwen3-8B 2026.05		6
Qwen3-32B 2026.05		6
Qwen2.5-32B 2026.05		5
Qwen3.5-35B 2026.05		4
GPT-4o 2026.05		0
GPT-5 2026.05		0
Claude-3.7 2026.05		0
Claude-4.6 2026.05		0
Qwen3-8B 2026.05		0
Qwen3-32B 2026.05		0
Qwen2.5-32B 2026.05		0
Qwen3.5-35B 2026.05		0