Open models and workplace agents lead new AI updates

European governments and businesses are adopting AI systems they can run on infrastructure they control. Projects in Poland, France and Germany are building public services on open models to keep sensitive data in-house. Mistral expanded Studio with managed agents and workflows that businesses can connect to their own data and run in their own cloud or on-premises. Mistral also released Mistral Large 4 as a downloadable open model for security work, reporting an 82% score on a test of fixing software vulnerabilities.

Creative and computing tools are expanding alongside those deployments. OpenAI announced Images 2.6 with faster image generation, GIF creation, more precise editing and new API controls across ChatGPT, Codex and the API. Google released Gemini Omni 1.2 Flash, saying it produces clips up to 15 seconds and lowers the price 30%. NVIDIA added StormScope EU, which predicts cloud changes across Europe every 10 minutes, with TotalEnergies evaluating solar-radiation nowcasts based on the model.

Workplace updates emphasize agent oversight and operational controls. Slack made its Code API generally available so developers can let agents open shared channels where teammates monitor, steer and approve work. GitHub expanded Copilot reporting to cover chat, agent activity and responses to code-review comments. OpenAI also said it would honor a shared robots.txt signal allowing website owners to state whether AI models may train on their pages.