ProductsOfficial
无问芯穹联合《三体》IP推出角色智能体，首 token 延迟降低 51.5%
Infinigence AI Builds a Three-Body Character Agent, Halving First-Token Time With Split Inference
Why we picked it
无问芯穹和三体宇宙合作做了个三体角色智能体，用 split inference 把首 token 延迟砍了一半多，做延迟敏感的 Agent 可以看看这个方案。
无问芯穹（Infinigence AI）与三体宇宙联合推出基于《三体》小说的角色智能体。该智能体运行在无问芯穹的 AgentWorks 平台上。官方称其采用异构 split inference 部署方案，首 token 生成时间降低了 51.5%。
Original · pandaily
Infinigence AI Builds a Three-Body Character Agent, Halving First-Token Time With Split Inference
Infinigence AI and Three-Body Universe launched a character agent from the novel. It runs on Infinigence's AgentWorks platform and a split inference setup that cuts first-token time by 51.5%.
