Long-form writing on EQ-Bench Long-form writing (leaderboard)
0.798ScoreClaude Sonnet 4.5
Evaluation Results
| Method | Links | |
|---|---|---|
| Claude Sonnet 4.5Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.798 | |
| ROMA w/ GEPA+Access Type=Open-source, Judge Model=Claude Sonnet 4, Backbone=DeepSeek-V3.12026.02 | 0.798 | |
| Claude Opus 4.5-2025-11-01Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.793 | |
| Polaris AlphaAccess Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.769 | |
| Claude Haiku 4.5Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.765 | |
| Qwen3-Max-2025-09-24Access Type=Open-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.751 | |
| Kimi K2.5Access Type=Open-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.749 | |
| GPT-5.2Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.745 | |
| Claude Opus 4.1Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.741 | |
| DeepSeek-V3.1Access Type=Open-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.736 | |
| Claude Sonnet 4Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.735 | |
| Kimi K2 ThinkingAccess Type=Open-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.731 | |
| DeepSeek V3.2Access Type=Open-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.726 | |
| Gemini 2.5 Pro Preview-03-25Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.72 | |
| Claude 3.7 Sonnet-2025-02-19Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.716 | |
| GPT-5-2025-08-07Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.714 | |
| Claude 3.5 SonnetAccess Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.709 | |
| DeepSeek-R1Access Type=Open-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.707 | |
| Gemini 3 Pro PreviewAccess Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.706 | |
| GPT-5 ChatAccess Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.706 | |
| Gemini 2.5 Pro Preview-06-05Access Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.706 | |
| Horizon AlphaAccess Type=Closed-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.7 | |
| GLM-4.6Access Type=Open-source, Judge Model=Claude Sonnet 4, From Leaderboard=true2026.02 | 0.7 |