Game Arena Werewolf
Teams of language models play the social deduction game Werewolf in natural language. Villagers must find the hidden werewolves, who try to deceive them.
Google DeepMind also presents it as a sandbox for deception and manipulation research. 8 models on the leaderboard in 2026-06.
- Made by
- Kaggle and Google DeepMind
- Released
- February 2026
- Kind
- Multi-agent games
- Level
- Expert
- Scored
- Ranked by matches
- Answers
- Agent task
- Tasks
- All public
- Strongest model tested
- GPT-5.4 #27 on the capability index
- Last model added
- March 2026
- Scores refreshed
- November 12th, 2026
GPT-5.4 leads with 0 among models released by March 2026
Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in March 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6 Astra
- Claude Opus 5.5
- and 16 more
- 1 GPT-5.4 0
- 2 Gemini 3.1 Pro -0
- 3 Gemini 3 Flash -1
- 4 Claude Opus 4.6 -1
- 5 Claude Sonnet 4.6 -2
- 6 GPT-5.4 Mini -6
- 7 Grok 4.1 Fast reasoning -6
- 8 Grok 4 -6
Source: Official Game Arena Werewolf leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.4 | 0 | Mar 2026 | Official leaderboard | |
| 2 | Gemini 3.1 Pro | -0 | Feb 2026 | Official leaderboard | |
| 3 | Gemini 3 Flash | -1 | Dec 2025 | Official leaderboard | |
| 4 | Claude Opus 4.6 | -1 | Feb 2026 | Official leaderboard | |
| 5 | Claude Sonnet 4.6 | -2 | Feb 2026 | Official leaderboard | |
| 6 | GPT-5.4 Mini | -6 | Mar 2026 | Official leaderboard | |
| 7 | Grok 4.1 Fast | -6 | reasoning | Nov 2025 | Official leaderboard |
| 8 | Grok 4 | -6 | Jul 2025 | Official leaderboard |
3 models have held number one on the official leaderboard since its release in February 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Game Arena Werewolf leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Game Arena Werewolf leaderboard
No lab model card reports it yet
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.
More in Games and long tasks
- ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6 Astra 99.95%
- Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Each game scores in-game progress or outcomes, and the board ranks models by their average rank across the games. Gemini 2.5 Pro 3.90
- MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks.
Seed2.0 Pro 49% - TextQuests Only in system cards Agents play 25 classic interactive fiction games. The score is harmful actions per 100 steps, where lower is better. The board also reports game progress.
Llama 4 Scout 7.00 - GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. GPT-6.1 Terra 76.8%
- Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%