模型多源确认精选72°

Synthetic Sciences 开源 OpenScience 智能体,科学基准超 Codex 和 Claude Code

精选理由

一个开源科研智能体把 Codex 和 Claude Code 都比下去了,Apache 2.0 免费随便用,还能接自家模型。

Synthetic Sciences 发布开源智能体工作台 OpenScience,采用 Apache 2.0 许可证,支持接入任意厂商的模型。在斯坦福和 Laude Institute 主办的 Terminal-Bench-Science 上,OpenScience 完成 70 个科研工作流中的 53 个,得分 75.7%,高于 Codex 搭载 GPT-6 Astra 的 68.1%。在 Terminal-Bench 4.0 的 14 个科学任务上,OpenScience 得分 71.4%,Claude Code 搭载 Claude Fable 5.1 为 60.0%。

原文 · rohanpaul_ai

Synthetic Sciences just launched its open-source OpenScience agent now outscores Codex and Claude Code on agentic science benchmarks.

A free Apache 2.0 workbench for any provider's model.

On Terminal-Bench-Science, 70 research workflows hosted by Stanford and the Laude Institute, OpenScience solved 53 tasks for 75.7%.

Codex with GPT-6 Astra scored 68.1%, the top entry on a September 23 mirror of the public leaderboard.

The gap widened on Terminal-Bench 4.0's 14 science tasks, where OpenScience scored 71.4% and Claude Code on Claude Fable 5.1 reached 60.0%.

GitHub link in comment