LMArena Agent Arena Bash Reliability
Agents
Signed Agent Arena diagnostic for successful bash commands, excluding failures outside the model's control. Atlas retains the publisher's raw bash_reliability_no_prereq sign and scales by 100; this baseline-relative signal can be negative and is not an absolute command-success probability.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 15.8 |
| Sonnet 5 | 15.4 |
| Claude Sonnet 5.5 | 13.2 |
| GPT-5.6 Sol | 12.0 |
| Claude Fable 5 | 11.9 |
| GPT-6 Sol | 11.7 |
| Claude Opus 5 | 11.3 |
| Gemini 3.8 Flash | 10.9 |
| GPT-6 Luna | 10.7 |
| Claude Fable 5.1 | 10.6 |
| Grok 4.7 | 9.7 |
| Grok 4.6 | 9.6 |
| GPT-6 Astra | 9.3 |
| Muse Wonder Full | 8.7 |
| Kimi K3.1 Pro | 8.3 |
| Claude Haiku 5.5 | 7.6 |
| Gemini 3.7 Flash | 6.8 |
| Muse Wonder | 6.5 |
| DeepSeek V4.1 Flash 1017 | 5.1 |
| Qwen3.7-Max (source unspecified) | 4.0 |
| GPT-6.1 Sol | 3.9 |
| MiMo-V2.6-Pro | 3.7 |
| Inkling-Small | 3.1 |
| DeepSeek V4 Pro 0813 | 2.4 |
| Muse Spark 1.3 | 2.3 |
| Gemini 3.1 Pro Preview | 2.3 |
| Qwen3.7-Plus (source unspecified) | 1.7 |
| GPT-5.6 Terra | 1.4 |
| Mistral Large 4 Preview | 1.1 |
| Inkling | 0.3 |
| MiniMax M3 | -0.1 |
| DeepSeek V4.1 Flash | -0.2 |
| MiMo-V2.5-Pro | -0.3 |
| Qwen3.8-Max | -0.9 |
| Muse Spark 1.1 | -1.4 |
| MiMo-V2.6-Flash | -1.9 |
| Hy4 Preview | -2.2 |
| Kimi K3 | -2.3 |
| Muse Spark 1.2 | -3.4 |
| GLM-5.2 | -4.0 |