Reported by others
Work in a terminal (Terminal Bench)
Tasks such as building programs, configuring systems and processing data, done by the model in a command-line environment.
How to read this table
- Higher rows are rated higher. A range like “1–2” means the gap is within the margin of error.
- The score is the value published by the task's authors and other organizations.
- Conditions can differ between models, so small gaps do not mean much.
- Prices are a guide per million tokens, reading then writing.
- Length is how much text the model can read at once (tokens).
± range of the top model
| Rank | Model | Developer | Success rate (%) | Input $ /1M | Output $ /1M | Context | Supports | Weights | Released | Price and specs |
|---|---|---|---|---|---|---|---|---|---|---|
| 1–5 | GPT-5.5 | OpenAI | Rating 84.7 ±4.1 | $5 | $30 | 1.05M | OpenAI Released 2026-04 Price $5 / $30 Length 1.05M | |||
| 1–6 | GPT-5.4 | OpenAI | Rating 81.8 ±3.9 | $2.5 | $15 | 1.05M | OpenAI Released 2026-03 Price $2.5 / $15 Length 1.05M | |||
| 1–6 | Claude Opus 4.7 | Anthropic | Rating 80.2 ±4.1 | $5 | $25 | 1M | Anthropic Released 2026-04 Price $5 / $25 Length 1M | |||
| 1–6 | Gemini 3.1 Pro (preview) | Rating 80.2 ±5.1 | $2 | $12 | 1.05M | Google Released 2026-02 Price $2 / $12 Length 1.05M | ||||
| 1–6 | Claude Opus 4.6 | Anthropic | Rating 79.8 ±3.1 | $5 | $25 | 1M | Anthropic Released 2026-02 Price $5 / $25 Length 1M | |||
| 2–6 | GPT-5.3 Codex | OpenAI | Rating 78.4 ±4.3 | $1.75 | $14 | 400K | OpenAI Released 2026-02 Price $1.75 / $14 Length 400K | |||
| 7–11 | Gemini 3 Pro (preview) | Rating 69.4 ±4.1 | — | — | 1M | Google Released 2025-11 Price — Length 1M | ||||
| 7–13 | GPT-5.2 Codex | OpenAI | Rating 66.5 ±6.0 | — | — | 400K | OpenAI Released 2025-12 Price — Length 400K | |||
| 7–13 | GPT-5.2 | OpenAI | Rating 64.9 ±5.4 | $1.75 | $14 | 400K | OpenAI Released 2025-12 Price $1.75 / $14 Length 400K | |||
| 7–13 | Gemini 3 Flash (preview) | Rating 64.3 ±5.5 | $0.5 | $3 | 1M | Google Released 2025-12 Price $0.5 / $3 Length 1M | ||||
| 7–13 | Claude Opus 4.5 | Anthropic | Rating 63.1 ±5.3 | $5 | $25 | 200K | Anthropic Released 2025-11 Price $5 / $25 Length 200K | |||
| 8–13 | GPT-5.1 Codex Mini | OpenAI | Rating 61.6 ±3.7 | — | — | 400K | OpenAI Released 2025-11 Price — Length 400K | |||
| 8–15 | GPT-5.1 Codex | OpenAI | Rating 60.4 ±5.2 | — | — | 400K | OpenAI Released 2025-11 Price — Length 400K | |||
| 13–16 | Grok 4.20 | xAI | Rating 57.3 | — | — | 2M | xAI Released 2026-02 Price — Length 2M | |||
| 13–19 | Claude Sonnet 4.6 | Anthropic | Rating 53.4 ±5.5 | $3 | $15 | 1M | Anthropic Released 2026-02 Price $3 / $15 Length 1M | |||
| 14–19 | GLM-5 | Z.ai | Rating 52.4 ±5.1 | $1 | $3.2 | 204.8K | Z.ai Released 2026-02 Price $1 / $3.2 Length 204.8K | |||
| 15–23 | GPT-5 | OpenAI | Rating 49.6 ±5.7 | $1.25 | $10 | 400K | OpenAI Released 2025-08 Price $1.25 / $10 Length 400K | |||
| 15–23 | GPT-5.1 | OpenAI | Rating 47.6 ±5.5 | $1.25 | $10 | 400K | OpenAI Released 2025-11 Price $1.25 / $10 Length 400K | |||
| 15–24 | Claude Sonnet 4.5 | Anthropic | Rating 46.5 ±4.6 | $3 | $15 | 200K | Anthropic Released 2025-09 Price $3 / $15 Length 200K | |||
| 17–23 | MiniMax-M2.7 | MiniMax | Rating 45.1 | $0.3 | $1.2 | 204.8K | MiniMax Released 2026-03 Price $0.3 / $1.2 Length 204.8K | |||
| 17–26 | GPT-5 Codex | OpenAI | Rating 44.3 ±5.3 | — | — | 400K | OpenAI Released 2025-09 Price — Length 400K | |||
| 17–28 | Kimi K2.5 | Moonshot | Rating 43.2 ±5.6 | — | — | 262.1K | Moonshot Released 2026-01 Price — Length 262.1K | |||
| 17–28 | MiniMax-M2.5 | MiniMax | Rating 42.7 ±5.5 | $0.3 | $1.2 | 204.8K | MiniMax Released 2026-02 Price $0.3 / $1.2 Length 204.8K | |||
| 20–31 | DeepSeek-V3.2 | DeepSeek | Rating 39.6 ±5.4 | — | — | 128K | DeepSeek Released 2025-12 Price — Length 128K | |||
| 21–31 | Claude Opus 4.1 | Anthropic | Rating 38.0 ±5.1 | — | — | 200K | Anthropic Released 2025-08 Price — Length 200K | |||
| 21–32 | MiniMax-M2.1 | MiniMax | Rating 36.6 ±5.7 | $0.3 | $1.2 | 204.8K | MiniMax Released 2025-12 Price $0.3 / $1.2 Length 204.8K | |||
| 22–32 | Kimi K2 (thinking) | Moonshot | Rating 35.7 ±5.5 | — | — | 262.1K | Moonshot Released 2025-11 Price — Length 262.1K | |||
| 22–32 | Claude Haiku 4.5 | Anthropic | Rating 35.5 ±5.7 | $1 | $5 | 200K | Anthropic Released 2025-10 Price $1 / $5 Length 200K | |||
| 24–34 | GPT-5 mini | OpenAI | Rating 34.8 ±5.3 | $0.25 | $2 | 400K | OpenAI Released 2025-08 Price $0.25 / $2 Length 400K | |||
| 24–36 | GLM-4.7 | Z.ai | Rating 33.4 ±5.6 | $0.6 | $2.2 | 204.8K | Z.ai Released 2025-12 Price $0.6 / $2.2 Length 204.8K | |||
| 24–37 | Gemini 2.5 Pro | Rating 32.6 ±5.9 | $1.25 | $10 | 1M | Google Released 2025-06 Price $1.25 / $10 Length 1M | ||||
| 26–38 | MiniMax-M2 | MiniMax | Rating 30.0 ±5.4 | $0.3 | $1.2 | 1M | MiniMax Released 2025-10 Price $0.3 / $1.2 Length 1M | |||
| 29–39 | Kimi K2 (instruct) | Moonshot | Rating 27.8 ±4.9 | — | — | 131.1K | Moonshot Released 2025-07 Price — Length 131.1K | |||
| 29–39 | Grok 4 | xAI | Rating 27.2 ±6.1 | — | — | 256K | xAI Released 2025-07 Price — Length 256K | |||
| 30–39 | Qwen3 Coder 480B A35B (instruct) | Alibaba | Rating 27.2 ±5.0 | $1.5 | $7.5 | 262.1K | Alibaba Released 2025-07 Price $1.5 / $7.5 Length 262.1K | |||
| 30–40 | Grok Code Fast 1 | xAI | Rating 25.8 ±5.1 | — | — | 256K | — | xAI Released 2025-08 Price — Length 256K | ||
| 31–42 | Qwen3.6 35B A3B | Alibaba | Rating 24.6 ±6.3 | $0.375 | $2.25 | 262.1K | Alibaba Released 2026-04 Price $0.375 / $2.25 Length 262.1K | |||
| 32–40 | GLM-4.6 | Z.ai | Rating 24.5 ±4.8 | $0.6 | $2.2 | 204.8K | Z.ai Released 2025-09 Price $0.6 / $2.2 Length 204.8K | |||
| 33–42 | GPT-5 nano | OpenAI | Rating 21.8 ±5.5 | $0.05 | $0.4 | 400K | OpenAI Released 2025-08 Price $0.05 / $0.4 Length 400K | |||
| 36–42 | GPT-OSS 120B | OpenAI | Rating 18.7 ±5.3 | — | — | 128K | OpenAI Released 2025-08 Price — Length 128K | |||
| 38–42 | Gemini 2.5 Flash (preview, 2025-09) | Rating 17.1 ±4.9 | — | — | 1M | Google Released 2025-09 Price — Length 1M | ||||
| 38–42 | Gemini 2.5 Flash | Rating 17.1 ±4.9 | $0.3 | $2.5 | 1M | Google Released 2025-06 Price $0.3 / $2.5 Length 1M | ||||
| 43 | Qwen3.5 9B | Alibaba | Rating 9.2 ±4.6 | — | — | 262K | Alibaba Released 2026-02 Price — Length 262K | |||
| 44 | GPT-OSS 20B | OpenAI | Rating 3.4 ±2.7 | — | — | 131K | OpenAI Released 2025-08 Price — Length 131K |
Best success rate over time
2025-06-17Release date2026-04-23
How to read this page
- Tests check the final state, and the value is the share of tasks completed. Figures published by the organizations that made the tasks, among others, collected here.
- The success rate is the share of tries that succeeded. A gap of 1 point is 1 success out of 100 tries.
- In the Supports column the icons are, from left: can use tools / can read images / can answer in a fixed format (JSON); faded means not supported or unknown.
- Open-weights models have their contents (the weights) published, so you can also run them on your own computer or server. Models without the open-weights mark are normally used through the developer's own services (apps or API); for some models the weights status is unknown.
- Names such as ChatGPT next to a model are the consumer services its developer offers. Check on the service itself whether that model is available there.
Data: Arena, Epoch AI, Swallow LLM team (CC BY 4.0), Models.dev (MIT), updated 2026-10-24, data sources