MECoBench - Parallel (1 Agent): leaderboard

Metric: Success rate (%) of a single agent in the parallel setting (every agent may enter every room; 60 total steps shared by the team), on 96 VirtualHome household rearrangement tasks (2 to 7 subgoals each) in which multimodal agents act from four-view first-person observations; printed 0-1 and x100 here; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 15 models tracked.

Top models

#ModelScore
1GPT-5.485.4
2Qwen 3.5 9B85.4
3Qwen 3.5 27B84.4
4Gemma 4 31B (IT)83.3
5Gemini 3.1 Pro (Preview)80.2
6Gemma 4 26B A4B (IT)76
7Qwen 3 VL 235B A22B Instruct65.6
8GLM-4.6V59.4
9Qwen 3 VL 32B Instruct58.3
10Qwen 3 VL 8B Instruct35.4
11Llama 4 Scout Instruct25

Interactive version: theaggregate.ai/benchmark?slug=mecobench-parallel-1-agent · How It Works · Data refreshed daily, snapshot 2026-09-29.