Event ·Official first-hand

SID AI 发布 SID-2 智能体搜索模型,Swarm 模式达 0.720 NDCG

SID AI 发布了用强化学习训练的智能体搜索模型 SID-2。单智能体在 3.11 秒内取得 0.585 NDCG;12 个智能体组成的 SID-2 Swarm 达到 0.720 NDCG,超过 GPT-6 Astra。 技术机制上,SID-2 通过强化学习训练执行搜索任务,并支持多智能体协作的 Swarm 模式,让多个 SID-2 实例并行检索后汇总结果,以此把成绩从单体的 0.585 提升到 12 智能体的 0.720,代价是时延从 3.11 秒增加。 需要注意,0.720 NDCG 是在 SID 自家基准上测得,对比对象 GPT-6 Astra 也在同一基准下评估,与其他公开基准没有直接可比性。对做检索增强(RAG)或搜索产品的开发者来说,多智能体换精度的思路值得对照自家场景权衡时延成本。

Current read

SID AI 把搜索做成了智能体:单个 3.11 秒拿 0.585 NDCG,12 个一起跑冲到 0.720,还压过 GPT-6 Astra。不过成绩是他们自家基准测的,看看多智能体并行搜索这个思路就好。

11 sources39° AI heatLast updated 2026/11/18 07:10:53

Evidence

11 sources
Primary sourcemarktechpost
View source
Related source 1arXiv cs.AI
View source
Related source 4Yuchen Jin
View source
Related source 5kimmonismus
View source
Related source 6lmarena.ai
View source
Related source 7TestingCatalog
View source
Related source 8PolymarketMoney
View source
Related source 9OpenRouter
View source
Related source 10Vaibhav Srivastav
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.