Third-party open models are now available on the Mistral AI platform, starting with Z.ai's GLM-5.2.



French AI company Mistral AI has announced that its platform 'Vibe' and other services will support not only its own AI models but also open models from third parties.

In-region inference, open models, and new European infrastructure for sovereign AI.

https://mistral.ai/news/regional-inference-open-models-new-compute/

Mistral AI is one of the largest AI companies in the EU, based in France, and offers products such as the 128 billion parameter AI model 'Mistral Medium 3.5,' the OCR model 'OCR 4' which also supports Japanese, and the speech synthesis model 'Voxtral TTS.'

French AI model 'Mistral Medium 3.5' surpasses Claude Sonnet 4.5 and is released as an open model & Mistral Vibe becomes cloud compatible - GIGAZINE



Japanese-language compatible OCR model 'Mistral OCR 4' released, AI converts PDFs and documents full of tables into 'usable data' - GIGAZINE



Mistral AI announces 'Voxtral TTS,' a text-to-speech AI model that allows users to clone their own voice. It supports 9 languages, boasts incredibly fast reading speeds, is lightweight, and is open-source. - GIGAZINE



These AI models are available for use in Mistral's integrated environment agent, 'Vibe,' and other applications.

Mistral AI rebrands Le Chat as 'Vibe,' creating an integrated environment for chat, coding, and job support agents - GIGAZINE



Mistral AI has now announced that it will begin supporting third-party open models to further expand its openness. This means that when using Vibe and other similar tools, users will no longer be limited to Mistral's AI models, and their model choices will be broader. Initially, Z.ai's GLM-5.2 will be supported.

GLM-5.2, a Chinese model surpassing Claude Opus 4.7, has been officially announced. It has even outperformed Claude Fable 5 in some tests and is available for download as an open model - GIGAZINE



There have already been reports of GLM-5.2 running on Vibe.



Additionally, we are launching the general availability of ' Mistral Regional Endpoints ,' which allows users to choose whether to run inference processing in the US or Europe based on their data storage location, regulatory requirements, and latency requirements.

Regional inference | Mistral Docs
https://docs.mistral.ai/inference/regional-inference

In addition, the ' Mistral Priority Tier ,' currently in public preview, offers guaranteed service levels for mission-critical workloads in AI labs, and is the only one in Europe to offer both 'processing region selection' and 'uptime guarantee.'

Priority Tier | Mistral Docs
https://docs.mistral.ai/inference/priority-tier

Related Posts:

in AI,   Web Service, Posted by logc_nt