Repository navigation
Biorouter v1.93.4
Biorouter v1.93.4
This patch release adds Baseten as a model provider. Baseten's Model APIs serve open-weight models such as GLM 5.3, Kimi K3, DeepSeek V4.1 and GPT OSS 120B, and Biorouter keeps their list and prices current between releases. Thank you to Albert Lee at Baseten, who contributed the provider.
Baseten
- Baseten is on the Public tab of the provider catalog, in Settings and on the first-run screen, and in
biorouter configure. It needs one API key,BASETEN_API_KEY. - Eleven models come with the app, with GLM 5.3 as the default: GLM 5.3, GLM 5.3 Fast, GLM 5.3 Flash, Kimi K3, two versions of DeepSeek V4.1 Flash, DeepSeek V4 Pro, Muse Spark, NVIDIA Nemotron 3.5 Super VL, NVIDIA Nemotron 3 Ultra and GPT OSS 120B.
- A chat's cost uses Baseten's prices, including the lower price for cached input, and the context gauge uses each model's own window.
- While a Baseten key is saved, Biorouter reads Baseten's published catalog at most twice a day. Prices stay current, and new Baseten models that can call tools appear in the model picker without an app update. The catalog is a public file on biorouter.ucsf.edu, so your key is never sent with that request, and without a Baseten key Biorouter makes no such request at all.
- Baseten is a public provider, so a private chat cannot switch to it.