Unverified50% confidenceFactExact time
Terminal Bench要求模型在交互式命令行环境中调试、运行、修复代码，区别于HumanEval等静态代码生成测试
1
Sources
50%
Confidence
Long-term
Relevance
7/8/2026
First Seen
Sources
GPT-5.6发布：Soul/Terra/Luna三层布局，编排能力训入模型
bilibili杨博士说AI6/29/2026
Related Claims
UnverifiedTerminal-Bench requires models to perform multi-turn interactions in a real shell environment, including issuing commands, parsing stdout/stderr output, handling permission issues, and managing filesystem state.70% similarVerifiedSWE-bench requires models to automatically generate patch code that passes tests given the full context of a code repository.67% similarUnverifiedTerminal Bench is an agent evaluation framework designed by Stanford University that simulates a real Linux terminal environment, requiring agents to complete full workflows including environment configuration, dependency installation, code debugging, and result verification.64% similarUnverified代理编程工具能够自主读写文件、调用终端命令、运行测试并根据结果迭代修改，整个过程几乎不需要人工干预每一步63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/248694API
curl https://kongchang.com/api/v1/knowledge/claims/248694MCP
get_claim(id=248694)