Skip to main content
Release Date: June 30, 2026
This release adds a new family of EU-hosted open-weight models you can pick right from the model menu, moves web search onto a GDPR-friendly European provider, and brings real per-model and per-provider breakdowns to the Analytics dashboard. Organization owners can now also choose which model powers behind-the-scenes tasks like chat titles, summaries, and document generation.

New EU-Hosted Models

Eight additional open-weight models are now available in the model picker — including GLM, DeepSeek, Qwen, Kimi, MiniMax and GPT-OSS — all served from EU infrastructure and grouped by data-residency region so you can clearly see which models stay in Europe.
The assistant’s web search now runs through a European, GDPR-compliant search provider, returning broader, more relevant results while keeping queries within EU data handling.

Richer Usage Analytics

The Analytics dashboard now shows real spend broken down by model and by provider — a model-spend-over-time chart, a cost-by-provider donut, and a per-model table — so you can see exactly where usage is going.

Choose Your Models for Internal Tasks

Organization owners get a new “AI Models” settings tab to override which model handles automatic tasks — chat-title generation, conversation summaries, image generation, and document creation — instead of always using the platform default.

Design Refresh Across Settings & Admin

A continued visual cleanup brings the refreshed, more consistent look to the settings, members, billing, profile, and admin screens.

More of the App in Your Language

Additional interface labels — including audit-log filters — are now fully translated, with German and English kept in sync.

Additional Improvements

  • Security & Maintenance: Routine updates to underlying libraries and infrastructure as part of ongoing security and maintenance.