CompassArena
Users chat with two anonymous models side by side, mostly in Chinese, and vote for the better answer. A Bradley-Terry model turns the votes into Elo-style ratings.
Launched on May 27th, 2024 with 23 models. The first board on June 13th, 2024 used about 6,000 battles. Later updates added multimodal battles and a search-off sub-board.
- Released
- May 2024
- Kind
- Overall chat quality
- Level
- Everyday
- Scored
- Ranked by votes
- Answers
- Open-ended
- Tasks
- Kept private
- Strongest model tested
- GPT-5.5 #20 on the capability index
- Last model added
- June 2026
- Scores refreshed
- November 12th, 2026
MiMo-V2.5-Pro leads with 1,062 among models released by June 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 June 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 15 more
Bars start at 960.
- 1 MiMo-V2.5-Pro 1,062
- 2
Qwen3.7-Max 1,055 - 3 Kimi K2.6 1,047
- 4
DeepSeek-V4-Pro 1,020 - 5 GLM-5.1 1,006
- 6 Claude Opus 4.8 995
- 7 Gemini 3.5 Flash 994
- 8
MiniMax-M3 990 - 9 GPT-5.5 989
- 10
Tencent Hy3 preview 963
Source: Official CompassArena leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | MiMo-V2.5-Pro | 1,062 | Apr 2026 | Official leaderboard | |
| 2 | Qwen3.7-Max | 1,055 | May 2026 | Official leaderboard | |
| 3 | Kimi K2.6 | 1,047 | Apr 2026 | Official leaderboard | |
| 4 | DeepSeek-V4-Pro | 1,020 | Apr 2026 | Official leaderboard | |
| 5 | GLM-5.1 | 1,006 | Apr 2026 | Official leaderboard | |
| 6 | Claude Opus 4.8 | 995 | May 2026 | Official leaderboard | |
| 7 | Gemini 3.5 Flash | 994 | May 2026 | Official leaderboard | |
| 8 | MiniMax-M3 | 990 | Jun 2026 | Official leaderboard | |
| 9 | GPT-5.5 | 989 | Apr 2026 | Official leaderboard | |
| 10 | Tencent Hy3 preview | 963 | Apr 2026 | Official leaderboard |
3 models have held number one on the official leaderboard since its release in May 2024
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official CompassArena 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 CompassArena 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 Instructions and chat
- IFEval In use Models answer about 500 prompts containing verifiable instructions such as word limits or required keywords. Code checks each instruction for strict and loose accuracy. Gemma 4 31B IT 98.9%
- IFBench In use 58 new out-of-domain verifiable output constraints. Scored by programmatic constraint checks (strict or loose).
Nemotron-Cascade 2 30B-A3B 82.9% - MultiChallenge In use Multi-turn conversations testing instruction retention, inference memory, versioned editing and self-coherence. Graded by LLM judge with rubrics. GPT-6.1 Sol 82%
- Multi-IF In use Models follow verifiable instructions over three-turn conversations in eight languages, extending IFEval. Scored by instruction accuracy at each turn. LFM3-1.7B 89.1%
- AdvancedIF In use Models respond to over 1,600 prompts with complex, multi-turn, and system-level instructions. A judge model checks each response against expert-written rubrics.
MAI-Thinking-1 85% - IFStruct In use Models generate JSON or YAML that must follow a requested schema and phrasing constraints, without constrained decoding. A response passes only if every structural constraint holds. Gemma 4 31B IT 95.9%