Show notes
I had to delay this a little bit, but the results are in and Kimi K2.5 is pretty good. Table of Contents Official Introduction. On Your Marks. Positive Reactions. Skeptical Reactions. Kimi Product Accounts. Agent Swarm. Who Are You? Export Controls Are Working. Where Are You Going? Safety Not Even Third. It's A Good Model, Sir. Official Introduction Introducing Kimi K2.5, Kimi.ai: Meet Kimi K2.5, Open-Source Visual Agentic Intelligence. Global SOTA on Agentic Benchmarks: HLE full set (50.2%), BrowseComp (74.9%) Open-source SOTA on Vision and Coding: MMMU Pro (78.5%), VideoMMMU (86.6%), SWE-bench Verified (76.8%) Code with Taste: turn chats, images & videos into aesthetic websites with expressive motion. Agent Swarm (Beta): self-directed agents working in parallel, at scale. Up to 100 sub-agents, 1,500 tool calls, 4.5× faster compared with single-agent setup. K2.5 is now live on http://kimi.com in chat mode and agent mode. K2.5 Agent Swarm in beta for high-tier users. For production-grade coding, you can pair K2.5 with Kimi Code. – API here. Tech blog here. Weights and code here. Wu Haoning (Kimi): We [...] ---Outline:(00:16) Official Introduction(03:16) On Your Marks(06:10) Positive Reactions(08:33) Skeptical Reactions(11:05) Kimi Product Accounts(11:39) Agent Swarm(13:06) Who Are You?(15:48) Export Controls Are Working(16:24) Where Are You Going?(19:47) Safety Not Even Third(20:55) It's A Good Model, Sir --- First published: February 4th, 2026 Source: https://www.lesswrong.com/posts/omSudRiFDvtNRrxZS/kimi-k2-5 --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.