Benchmark
Terminal-Bench 2.0
Terminal-Bench 2.0 is an updated benchmark for testing AI agents' tool use ability to operate a computer via terminal. It evaluates how well models can…
- Modality
- text
- Categories
- reasoning, agents, code, tool_calling
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 51
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Claude Mythos Preview | Anthropic | 0.82 | 2026-04-07 | evidence |
| Claude Opus 4.5 | Anthropic | 0.593 | 2025-11-24 | evidence |
| Claude Opus 4.6 | Anthropic | 0.654 | 2026-02-05 | evidence |
| Claude Opus 4.7 | Anthropic | 0.694 | 2026-04-16 | evidence |
| Claude Opus 4.8 | Anthropic | 0.746 | 2026-05-28 | evidence |
| Claude Sonnet 4.6 | Anthropic | 0.591 | 2026-02-17 | evidence |
| Claude Sonnet 5 | Anthropic | 0.804 | 2026-06-30 | evidence |
| DeepSeek-V3.2 | DeepSeek | 0.464 | 2025-12-01 | evidence |
| DeepSeek-V3.2 (Thinking) | DeepSeek | 0.464 | 2025-12-01 | evidence |
| DeepSeek-V3.2-Speciale | DeepSeek | 0.464 | 2025-12-01 | evidence |
| DeepSeek-V4-Flash-0423 | DeepSeek | 0.566 | 2026-04-23 | evidence |
| DeepSeek-V4-Flash-Max | DeepSeek | 0.569 | 2026-04-23 | evidence |
| DeepSeek-V4-Pro-Max | DeepSeek | 0.679 | 2026-04-23 | evidence |
| Gemini 3 Flash | 0.476 | 2025-12-17 | evidence | |
| Gemini 3 Pro | 0.542 | 2025-11-18 | evidence | |
| Gemini 3.1 Pro | 0.685 | 2026-02-19 | evidence | |
| Gemini 3.5 Flash | 0.762 | 2026-05-19 | evidence | |
| GLM-4.7 | Z.ai | 0.41 | 2025-12-22 | evidence |
| GLM-5 | Z.ai | 0.562 | 2026-02-11 | evidence |
| GLM-5.1 | Z.ai | 0.69 | 2026-04-07 | evidence |
| GPT-5.1 Codex | OpenAI | 0.528 | 2025-11-19 | evidence |
| GPT-5.2 Codex | OpenAI | 0.64 | 2026-01-14 | evidence |
| GPT-5.3 Codex | OpenAI | 0.773 | 2026-02-05 | evidence |
| GPT-5.4 | OpenAI | 0.751 | 2026-03-05 | evidence |
| GPT-5.4 mini | OpenAI | 0.6 | 2026-03-17 | evidence |
| GPT-5.4 nano | OpenAI | 0.463 | 2026-03-17 | evidence |
| GPT-5.5 | OpenAI | 0.827 | 2026-04-23 | evidence |
| Kimi K2.5 | Moonshot AI | 0.508 | 2026-01-27 | evidence |
| Kimi K2.6 | Moonshot AI | 0.667 | 2026-04-20 | evidence |
| Laguna XS 2.1 | Poolside | 0.375 | 2026-07-02 | evidence |
| MAI-Code-1-Flash | Microsoft | 0.548 | 2026-06-02 | evidence |
| MAI-Thinking-1 | Microsoft | 0.46 | 2026-06-02 | evidence |
| MiMo-V2-Flash | Xiaomi | 0.385 | 2025-12-16 | evidence |
| MiMo-V2-Pro | Xiaomi | 0.571 | 2026-03-18 | evidence |
| MiMo-V2.5 | Xiaomi | 0.658 | 2026-04-22 | evidence |
| MiMo-V2.5-Pro | Xiaomi | 0.684 | 2026-04-27 | evidence |
| MiniMax M2.7 | MiniMax | 0.57 | 2026-03-18 | evidence |
| Muse Spark | Meta | 0.59 | 2026-04-08 | evidence |
| Nemotron 3 Super (120B A12B) | NVIDIA | 0.31 | 2026-03-11 | evidence |
| North Mini Code 1.0 | Cohere | 0.36 | 2026-06-09 | evidence |
| Qwen3-Coder 480B A35B Instruct | Qwen | 0.375 | 2025-01-31 | evidence |
| Qwen3.5-122B-A10B | Qwen | 0.494 | 2026-02-24 | evidence |
| Qwen3.5-27B | Qwen | 0.416 | 2026-02-24 | evidence |
| Qwen3.5-35B-A3B | Qwen | 0.405 | 2026-02-24 | evidence |
| Qwen3.5-397B-A17B | Qwen | 0.525 | 2026-02-16 | evidence |
| Qwen3.6 Plus | Qwen | 0.616 | 2026-04-02 | evidence |
| Qwen3.6-27B | Qwen | 0.593 | 2026-04-21 | evidence |
| Qwen3.6-35B-A3B | Qwen | 0.515 | 2026-04-16 | evidence |
| Qwen3.7 Max | Qwen | 0.697 | 2026-05-19 | evidence |
| Qwen3.7-Plus | Qwen | 0.703 | 2026-05-31 | evidence |
| Step-3.5-Flash | StepFun | 0.51 | 2026-02-02 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.