Meta 的自主科研系统 AIRA₃ 在 NVIDIA Kaggle 竞赛中获金牌
As a test of our progress to advance the frontier of AI research, in June we entered the next generation of our autonomous AI research system, AIRA₃, in a live
Meta 在 6 月让下一代自主 AI 研究系统 AIRA₃ 参加 NVIDIA 在 Kaggle 举办的竞赛，任务是微调 30B Nemotron 模型以提升其推理能力，所有参赛者信息相同并在私有测试集上外部评分。AIRA₃ 在约 4000 支队伍中排名第 8，获得金牌，超过使用相同前沿工具的人类参赛者。AIRA₃ 不依赖中央控制器，而是让大量长时运行的智能体在隔离环境中通过共享论坛和共享文件系统异步协作；其金牌方案由 GPT 5.5（配 OpenCode）与 Claude 4.8（配 ClaudeCode）组合而成，事后评估中 Muse Spark 1.2（配 MuseCode）也达到金牌水平，Muse Spark 1.1（配 OpenCode）和 GLM 5.2（配 OpenCode）达到银牌水平。Meta 称该系统只改任务说明即可跨领域泛化，内部基准上实现生产 GPU kernel 延迟降低 27%，并在另一场将 4000 年前阿卡德语泥板翻译成英文的 Kaggle 竞赛中取得金牌级表现。
推荐理由
精选AI 评分 63
Meta 公开了 AIRA₃ 在 Kaggle 微调竞赛中的名次与多智能体协作机制，可据此观察自动化科研系统的当前水平。
主题
正文
当前显示原文，中文全文译文尚未就绪。
As a test of our progress to advance the frontier of AI research, in June we entered the next generation of our autonomous AI research system, AIRA₃, in a live Kaggle competition run by NVIDIA to fine-tune a 30B Nemotron model. The challenge was to teach the model to reason better — all competitors had access to the same information and were graded externally on a private test set.
AIRA₃ placed 8th out of ~4,000 teams to win Gold, outperforming human competitors who had access to the same frontier tools.
We believe this is a reliable signal that AIRA₃ can improve a targeted capability of an AI model at a level similar to human experts.