CogEvol：面向高效可靠的学习环境生成

摘要

我们提出CogEvol，这是一个专门针对学习环境生成（Learning Environment Generation）任务训练的模型系列：将课程简报一次性转化为完整的学习产物（结构化JSON幻灯片或自包含的交互式HTML页面）。在22万次生产请求中，CogEvol生成一张幻灯片的中位耗时为17秒，生成一个交互式页面的中位耗时为59秒，取代了此前耗时数分钟的多轮智能体脚手架流程。可靠性不是寄望于偶然，而是通过机制加以保障：一条基于生产数据的流水线将真实失败案例转化为53，687个经过验证的SFT样本，混合规则加VLM奖励驱动基于GRPO的强化学习——我们在发现并修复了一次产生视觉上可信但无法游玩的游戏的奖励黑客事件后，对该流程进行了加固。CogEvol-27B在幻灯片质量上取得83.7分，在500个案例的交互式HTML基准测试上取得63.7分，参数量仅为旗舰级编程模型的二十六点九分之一，并与OpenMAIC团队合作，服务于其生产环境的实时流量。CogEvol-4B已在Apache 2.0许可证下开源，地址为https://github.com/CogEvol/CogEvol-4B；外部旗舰模型在相同测试套件和完全相同的评测框架下进行测量。脚手架编辑可将交互式页面的生成成本进一步降低约76%，且整个技术栈可在国产昇腾加速器上运行，应用层面的性能与A800 GPU持平，从而在大规模场景下降低AI原生教育的单位成本。

English

We present CogEvol, a family of models trained specifically for Learning Environment Generation: turning a course brief into a finished learning artifact (structured-JSON slides or self-contained interactive HTML pages) in a single pass. Across 220k production requests, CogEvol completes a slide in a median of 17 seconds and an interactive page in 59, replacing minutes-long multi-turn agent scaffolding. Reliability is enforced rather than hoped for: a production-grounded data pipeline turns real failures into 53,687 verified SFT samples, and a hybrid rule-plus-VLM reward drives GRPO-based RL, hardened after we caught and fixed a reward-hacking episode that produced visually convincing but unplayable games. CogEvol-27B scores 83.7 on slide quality and 63.7 on a 500-case interactive-HTML benchmark with 26.9x fewer parameters than flagship coding models, and, in collaboration with the OpenMAIC team, serves their live production traffic. CogEvol-4B is released openly under the Apache 2.0 license at https://github.com/CogEvol/CogEvol-4B; external flagships are measured on the same suites under the identical harness. Scaffold editing cuts interactive-page generation cost by a further ~76%, and the full stack runs on domestic Ascend accelerators at application-level parity with A800 GPUs, lowering the unit cost of AI-native education at scale.