Subscribe
Mistral
2 min read

Mistral Expands Platform to Host Third-Party Open Weights Starting with GLM-5.2

Mistral's platform now hosts third-party open models, starting with Z.ai's GLM-5.2, on the same regional controls as its own models. It also made Regional Endpoints generally available, previewed a Priority Tier with an SLA, and announced European Compute Units for long-term capacity.

Mistral Expands Platform to Host Third-Party Open Weights Starting with GLM-5.2

Update: Mistral's docs now list zai-glm-5-2 as deprecated as of September 29, 2026, with Z.ai GLM 5.3 as the replacement for new integrations. See our GLM-5.3 coverage.

Mistral AI said on August 11, 2026 that its platform will host third-party open models, starting with Z.ai's GLM-5.2. Until now, Mistral only ran Mistral models. The company says GLM-5.2 and future open models will run on the same infrastructure, regional controls and service commitments as its own models.

Mistral's model card describes GLM-5.2 as a third-party open weight text model for long-context coding and agentic workflows, served without Mistral modifications, with a 1 million token context window and a 128,000 token maximum output. Its API identifier is zai-glm-5-2.

Mistral Sovereign Infrastructure and Multi-Model Hosting Architecture

Regional Endpoints and Priority Tier

The model hosting was one of three steps Mistral announced. The other two cover where inference runs and how reliable it is:

  1. Regional Endpoints, now generally available: customers choose whether inference runs in Europe or the US. Inference and the associated processing take place in the selected region, subject to limited, safeguarded transfers to sub-processors that may occur outside it, as described in Mistral's Trust Center.
  2. Priority Tier, in public preview: committed service levels for mission-critical workloads, including custom rate limits and an uptime SLA.

Mistral says it is the only European AI lab to offer both a choice of processing region and an SLA-backed service level.

Why Mistral is adding other models

Mistral says AI systems are now built on "an ensemble of capabilities" rather than a single model. It describes open weights as giving customers "the ability to see inside a model, adapt it, and retain the intelligence they build with it," and says it contributes to the Open Secure AI Alliance and the NVIDIA Nemotron Coalition.

"Mistral allows us to run open models under strict regional controls and service commitments, making it easy for us to maintain data residency and compliance requirements," said Matan Grinberg, CEO and cofounder of Factory.

European Compute Units

The third step is a coalition for long-term compute in Europe. Mistral is bringing together an anchor group of enterprises whose multi-year commitments can support infrastructure at a scale no single participant could secure. European Compute Units, or ECUs, turn those commitments into multi-year access to Mistral-built infrastructure on Mistral Compute. The announcement included supporting statements from the CEOs of Amadeus, ASML, Capgemini, Caisse des Dépôts and CMA CGM. Mistral plans to build up to 1 GW of capacity by 2030.

Mistral argues that "without assured access to that compute, Europe cannot control the AI systems on which its industries, institutions, and future competitiveness will depend."

Sources

MistralOpen WeightsInferenceInfrastructure