Deprecation of Leanstral 1.5
Leanstral 1.5 (labs-leanstral-1-5), an experimental Labs model, is deprecated and retires on September 30, 2026.
Also shipped Sep 29, 2026Mistral AI in September 2026 →
Horizontal scaling for Databricks Apps is now generally available
Horizontal scaling for Databricks Apps is now generally available. You can run a Databricks app across multiple instances behind one app URL for higher availability and concurrency, with zero-downtime deployments and best-effort session affinity. See Horizontal scaling for Databricks apps .
Elastic Cloud Serverless: Another Azure region
Elastic Cloud Serverless is now available in the Microsoft Azure Iowa region. This expansion brings the total number of supported serverless Azure regions to 12.
New Relic Control host-based fleets are now generally available
Linux and Windows host support for Agent Control and Fleet Control reaches General Availability today, matching the GA support for Kubernetes that shipped in September 2025. New Relic Control now gives you one observability control plane across Kubernetes, Linux, and Windows hosts, so your entire fleet is managed the same way no matter where it runs.
Triage for GitLab | Cloud and Self-Hosted | Team, Beta
Triage for GitLab | is now available in beta for GitLab Cloud and self-managed GitLab organizations on the Team plan and higher. The Triage documentation covers access requirements, prioritization, views, and supported actions.
Rich link previews for Agent Runners sites
Every site created with Agent Runners now includes Open Graph tags. When you share a link to your site, it shows up as a proper preview instead of a bare URL. Social media sites and messaging apps read Open Graph tags to build link previews.
More from Mistral AI Studio and APIEvery Mistral AI Studio and API release →
General Availability of Z.ai GLM 5.3
Z.ai GLM 5.3 (zai-glm-5-3) is now Generally Available.
Introducing Mistral Large 4
Mistral AI has released Mistral Large 4 via its API and Studio.
Mistral and Mozilla are bringing open, private and multilingual AI to your web browser
Open, private and multilingual AI is coming to your web browser. Mistral and Mozilla team up to put powerful, trustworthy AI where you already browse.
Introducing Timestral
Mistral AI introduced Timestral, a family of zero-shot probabilistic forecasting models ranging from 5M to 209M parameters. The CPU-efficient models are designed for time-series forecasting tasks.
Voxtral 3
Voxtral 3 is a voice agent SDK that supports real-time speech in over 50 languages. It features accurate transcription, emotional text-to-speech, and tool use capabilities.
Voxtral 3 voice agent SDK and speech models
Voxtral 3 introduces a voice agent SDK and speech models for real-time, multi-step conversations in over 50 languages. It supports transcription, natural speech synthesis with emotional adaptability, and tool integration for workflow automation.
Voxtral Mini Transcribe 3
Voxtral Mini Transcribe 3 is a speech-to-text model designed for multilingual transcription and diarization. It supports over 50 languages and context biasing for improved accuracy on domain-specific vocabulary.
Mistral and NVIDIA collaboration for engineering AI
Mistral and NVIDIA are partnering to integrate AI models with accelerated computing and physics modeling tools. The collaboration focuses on developing public benchmarks and open datasets for physics and agentic AI in computer-aided engineering.