About Arena-Hard v2
Definition and scoring
- Organisation
- LMSYS
- Category
- Instruction Following
- Version
- v2
- Direction
- higher is better
- Ranking use
- Reference
- Contamination risk
- Medium
Hard preference-style chat evaluation derived from Arena battle prompts, measuring instruction-following and helpfulness under automated judging. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗