Terminal-Bench 4.0 — Hardware
Agents · 2026-09-22
Resolution rate on the five hardware tasks in Terminal-Bench 4.0.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra | 73.3 |
| Gemini 4 Carbon | 60.0 |
| Gemini 4 Argon | 46.7 |
| Muse Wonder Full | 46.7 |
| Muse Wonder | 46.7 |
| GPT-6.1 Sol | 40.0 |
| Kimi K3.1 Pro | 40.0 |
| MiMo-V2.6-Pro | 33.3 |
| Opus 4.8 | 26.7 |
| Gemini 3.8 Flash | 26.7 |
| GPT-6.1 Terra | 26.7 |
| Grok 4.7 | 26.7 |
| Claude Haiku 5.5 | 20.0 |
| GLM-5.3 Flash | 20.0 |
| GLM-5.3 | 20.0 |
| GPT-5.6 Sol | 20.0 |
| Step 5 Preview | 20.0 |
| Gemini 3.7 Flash | 13.3 |
| Kimi K3 | 13.3 |
| MiMo-V2.6-Flash | 13.3 |
| Mistral Large 4 Preview | 13.3 |
| Sonnet 5 | 6.7 |
| DeepSeek V4.1 Flash 1017 | 6.7 |
| DeepSeek-V4-Flash-0731 | 6.7 |
| Ember-1 | 6.7 |
| Gemini 3.5 Flash | 6.7 |
| GPT-5.6 Terra | 6.7 |
| GPT-6 Sol | 6.7 |
| Grok 4.6 | 6.7 |
| MiniMax M3.1 Flash | 6.7 |
| Qwen3.8 Max (0902) | 6.7 |
| Beam 1.0 501B A23B | 0.0 |
| DeepSeek V4.1 Flash | 0.0 |
| DeepSeek V4 Pro 0813 | 0.0 |
| Gemini 3.1 Pro Preview | 0.0 |
| GPT-5.4 Mini | 0.0 |
| GPT-5.6 Luna | 0.0 |
| GPT-6 Luna | 0.0 |
| Grok 4.5 | 0.0 |
| Hy4 Preview | 0.0 |