概念
也叫：terminalbench
斯坦福大学开发的 AI 模型终端任务性能评估基准。
最近变化
2026-06-27 · GPT 5.6 Soul 在终端任务中超越 Claude Mythos 5。
Terminal Bench 被反复提及时，通常意味着它正在影响产品路线、开发者工作流或 AI 产业判断。这个页面把分散材料合并成一个可持续更新的观察入口。
OpenAI Just Introduced GPT 5.6 (Beats Claude Fable 5 And Mythos)
TheAIGRID · 8.5 分
GLM 5.2 is live on Fireworks, day zero. 1M-token context, coding‑first frontier model, independently...
Fireworks AI(@FireworksAI_HQ) · 8.5 分
1/ Today at #GoogleIO, we’re releasing Gemini 3.5, our latest family of models combining frontier in...
Jeff Dean(@JeffDean) · 8.5 分
已收录 5 篇与「Terminal Bench」相关的 AI 资讯和分析。
OpenAI 推出 GPT 5.6 系列模型，包含 Soul、Terror 和 Luna，Soul 超越 Claude Mythos 5 在终端任务表现。
入选理由：GPT 5.6 Soul 在终端任务中超越 Claude Mythos 5。
Fireworks AI 已上线 GLM 5.2 模型，支持 1M-token 上下文，专注于代码生成，并在多个基准测试中表现优异。
入选理由：GLM 5.2 支持 1M-token 上下文，适用于复杂任务。
Google releases the Gemini 3.5 family, starting with 3.5 Flash for complex agentic workflows. It outperforms 3.1 Pro on coding and agent benchmarks and runs 4x faster, reaching 12x in Antigravity.
入选理由：Gemini 3.5 Flash 专为执行复杂、长周期的智能体工作流而设计。
Thomas Wolf is excited about the extension of Terminal-Bench to scientific fields, known as Terminal-Bench Science. This benchmark evaluates AI models' ability to control tools via the command line to achieve scientific goals. It's open for contributions of real scientific workflows until August 2026, aiming to improve AI models' assistance in research work.
入选理由：Terminal-Bench Science evaluates AI models' performance in handling scientific workflows through command-line tools.
A harness is the core infrastructure for building AI Agents, consisting of tools, execution environments, system prompts, and file systems. By optimizing harness engineering, developers can significantly boost Agent performance on benchmarks like Terminal Bench without changing the underlying model.
入选理由：Harness 定义为模型访问的工具、执行环境、系统提示词和文件系统的集合。
与「Terminal Bench」经常一起出现的 AI 术语。
💡 想追踪「Terminal Bench」的长期趋势？去 实体雷达 · Terminal Bench 查看详细分析和跨材料问答。
