LMArena Agent Arena Chat Confirmed Success
Agents · 2026-08-14
Signed percentage-point Agent Arena Chat category diagnostic signal for sessions where the user explicitly confirmed task completion, recomputed on sessions tagged as general agentic chat such as creative writing, learning, everyday research and media generation. Values may be negative relative to the category baseline; reported confidence intervals are not standard errors.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 31.1 |
| Claude Sonnet 5.5 | 24.7 |
| Gemini 4 Argon | 19.7 |
| GPT-6 Astra | 18.8 |
| Claude Fable 5.1 | 18.2 |
| Muse Wonder Full | 14.1 |
| Claude Opus 5 | 12.8 |
| GPT-6.1 Sol | 12.5 |
| Grok 4.7 | 12.1 |
| Gemini 3.8 Flash | 11.0 |
| MiMo-V2.6-Pro | 8.5 |
| GPT-6 Sol | 8.3 |
| DeepSeek V4.1 Flash 1017 | 7.1 |
| Muse Wonder | 5.9 |
| Claude Fable 5 | 5.8 |
| Hy4 Preview | 5.3 |
| DeepSeek V4.1 Flash | 4.9 |
| Kimi K3 | 4.4 |
| Qwen3.8-Max | 4.2 |
| MiMo-V2.6-Flash | 4.2 |
| Gemini 3.7 Flash | 3.6 |
| GLM-5.3 Flash | 3.1 |
| Muse Spark 1.3 | 3.1 |
| GPT-5.6 Sol | 2.4 |
| Kimi K3.1 Pro | 1.1 |
| Mistral Large 4 Preview | 0.8 |
| Muse Spark 1.1 | 0.6 |
| GPT-6 Luna | -0.3 |
| GLM-5.3 | -0.5 |
| Qwen3.8-Flash-Next | -0.6 |
| Claude Haiku 5.5 | -2.3 |
| Grok 4.6 | -3.2 |
| Qwen3.8 27B | -3.4 |
| GPT-5.6 Terra | -3.5 |
| GLM-5.2 | -3.6 |
| Muse Spark 1.2 | -4.8 |
| Sonnet 5 | -4.9 |
| Step 5 Preview | -5.2 |
| Qwen3.7-Plus (source unspecified) | -6.6 |
| DeepSeek V4 Pro 0813 | -8.9 |