Mistral says its platform will support third-party open models, beginning with Z.ai’s GLM-5.2, according to a Techmeme item summarizing a Mistral AI Blog post. The company says those third-party models will run on the same infrastructure as Mistral’s own models. That is the concrete shift in the announcement: Mistral is positioning its platform not only as a distribution and deployment layer for its own models, but also as a place where outside open models can be run. The first named model is GLM-5.2 from Z.ai. The feed summary does not include additional detail on model capabilities, pricing, regional availability, service-level terms, or whether support is generally available immediately. Mistral’s own framing, as visible in the summary, centers on control: the blog begins by saying the company believes every enterprise and country must be in control of the models it uses. The full wording is truncated in the feed item, so the broader policy or product argument should not be overstated from the available material. For buyers, the relevant point is optionality. If Mistral follows through, customers using its platform could have a path to run at least some third-party open models through the same infrastructure stack they use for Mistral models. For now, the confirmed detail is narrow: third-party open-model support is being introduced, and GLM-5.2 is the starting example. Who benefits: Mistral benefits if customers see its platform as a neutral place to run multiple open models. Z.ai also gains distribution if GLM-5.2 becomes available through Mistral’s infrastructure. Who's exposed: Model platforms that rely on exclusive or tightly controlled model catalogs could face comparison if customers prefer broader open-model choice. The feed summary is too thin to identify specific competitors or commercial impact.