MiMo V2.6, GLM 5.3 FlashX, Qwen 3.8 and Muse Spark 1.3 join the catalog, so each Mate can run on the model that best fits its job, open-weight options included.

The big three labs are not the only ones shipping. Eight alternative models join the catalog, so your teams can pick the right model for each Mate, including open-weight ones.
Mates on Qwen 3.7 Flash now run on Qwen 3.8 Flash, and Mates on Muse Spark 1.2 on Muse Spark 1.3. Every other Mate keeps its model: to try a new one, open the Mate’s settings and pick it in the model list.
A 30-minute walkthrough of this update, based on your own use cases.
Talk to an expert
Our team will show you how governed AI teammates take on your most critical processes, from first use case to enterprise scale.