CoCoBench: leaderboard
CoCoBench (2026) evaluates multimodal LLMs as centralized planners for two to four cooperating embodied agents on 897 executable household tasks. Task success rate is the share of episodes whose goal predicates all hold at the end, in percent; higher is better.
Metric: Task success rate (%; share of episodes whose goal predicates all hold at the end; 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: Around February 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 84.8 | #29 |
| 2 | Claude Opus 4.8 | 78.5 | #56 |
| 3 | Qwen 3.6 Plus | 69.1 | #217 |
| 4 | Claude Haiku 4.5 | 63.1 | #568 |
| 5 | Kimi K2.6 | 57.6 | #173 |
| 6 | Qwen 3.5 9B | 36.7 | #803 |
| 7 | GPT-5.4 Mini | 33.6 | #389 |
| 8 | Qwen 3 VL 8B Instruct | 24.6 | #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 · How It Works · Data refreshed daily, snapshot 2026-11-12.