CoCoBench - Mutual Exclusion: leaderboard
CoCoBench (2026) evaluates multimodal LLMs as centralized planners for two to four cooperating embodied agents on 897 executable household tasks. Task success rate in percent, higher is better, on the mutual-exclusion instances, where agents share a single tool without blocking each other.
Metric: Task success rate (%; mutual-exclusion instances, sharing a tool or resource without contention; 897 oracle-validated executable household multi-agent task instances with two to four agents, centralized planner with image observations, task-specific step budgets). Source: arxiv.org. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 96.5 | #29 |
| 2 | Claude Opus 4.8 | 93.9 | #56 |
| 3 | Kimi K2.6 | 93.5 | #173 |
| 4 | Qwen 3.6 Plus | 88.3 | #217 |
| 5 | Claude Haiku 4.5 | 75.7 | #568 |
| 6 | Qwen 3.5 9B | 17 | #803 |
| 7 | GPT-5.4 Mini | 10.9 | #389 |
| 8 | Qwen 3 VL 8B Instruct | 1.3 | #982 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=cocobench-mutual-exclusion · How It Works · Data refreshed daily, snapshot 2026-11-12.