Changelog
New features, improvements and updates at a glance.
New Frontier Models and More Reliable Image Analysis
GPT-5.6 Luna, Terra, and Sol
The new GPT-5.6 family expands the model selection for different tasks: Luna for fast everyday requests, Terra for versatile business tasks, and Sol for particularly demanding analysis and reasoning.
Claude Sonnet 5
Claude Sonnet 5 replaces the previous Sonnet generation and is available for writing, coding, image understanding, and more complex tasks.
Updated model recommendations
Recommendations for marketing, software development, analysis, everyday questions, and document and image review now include the new models and match them more closely to each task.
Claude image analysis stabilized
Images are now transferred to Claude models in the appropriate format. This makes analysis of uploaded images with Claude Opus and Sonnet more reliable.
Research, code understanding and image editing improved
New and updated AI models
Kimi K2.7 Code is now available as a powerful EU-provider model. GLM 5.2 and DeepSeek V4 Pro have been updated for larger contexts and are ready for demanding chat and analysis tasks.
Semantic search for better code and knowledge context
Lurus can now search content more by meaning instead of only exact terms. This helps surface relevant project information and technical relationships faster.
Fetch web content directly
Current webpages and documentation can be included more precisely, so answers for research and implementation tasks are grounded in up-to-date sources.
New image editing
Image editing has been expanded so existing images can be adjusted more precisely and new variants can be generated from an existing visual.
Faster responses with optimized models
Open-source models now use better-suited model variants for recurring context. This improves response time and stability during longer workflows.
More reliable streaming
Chat responses are streamed more robustly, and ongoing answers remain visible more reliably during longer generations.
Chat status messages stabilized
Repeated or incorrectly placed status messages during running answers have been reduced, making the chat feel calmer and easier to follow.
Image editing fixes
Several issues around image editing and image variants have been fixed.
Lurus v2 – new AI architecture, MCP and memories
New model portfolio with auto selection
Lurus combines EU provider models with frontier models from Anthropic, Google and OpenAI and selects suitable capabilities for chat, analysis, research and generation tasks on every request – including very large context windows up to 1M tokens on supported models.
Configurable reasoning effort
Reasoning effort can now be tuned more precisely – from faster responses to deeper analysis for complex tasks.
Default models per task
You can define default models for image generation, document analysis, image analysis and context summarization.
MCP servers for custom tools
External tool servers can now be connected via MCP, managed in Lurus and used in controlled AI workflows with approval steps.
Personalization and memories
Response style, tone, custom instructions and saved memories can be managed directly in settings – visible, editable and removable.
Lurus Boost for frontier models
Lurus Boost scales token quotas and plans by a factor of 5, making demanding frontier models practical for larger workflows.
More stable context for long conversations
The new context architecture keeps long chats more reliable, summarizes conversations when needed and better preserves important working state.
More robust tool and integration workflows
Tool executions, status updates and confirmations are handled in a more structured way so local and remote chats behave more consistently.
More granular controls for teams
Teams gain more control over model usage, default models, integrations and memory features.
Minor bug fixes
Minor bug fixes were made.
New AI Models & Benchmark Comparison
New AI Models
Kimi K2.7 Code, GLM 5.2, DeepSeek V4 Pro, Qwen 3.5 397B, and frontier models from Anthropic, Google, and OpenAI are available in the curated portfolio.
Benchmark Comparison
New comparison page: See how EU-provider models and frontier models are positioned by task, privacy need, and performance profile.
Performance & Stability
General performance optimizations and improved stability.
Bug Fixes
Minor bug fixes and UI corrections.
Document Generation, Security & Stability
Document Generation
New feature for direct file generation by AI: Create documents, spreadsheets and more – directly from the chat.
Enhanced Prompt Injection Protection
Improved protection against prompt injection attacks on tool calls.
Performance & Stability
General performance optimizations: faster load times and improved stability.
Bug Fixes
Bug fixes and minor UI corrections.
Invoice Management, Deep Research & Stability
Invoice Management
Full invoice management directly in Lurus: overview of all invoices, PDF download, and transparent cost breakdown per billing period.
Deep Research
In-depth research with AI support: Lurus searches multiple sources, analyzes content, and creates comprehensive reports with citations – directly in chat.
Performance & Stability
General performance optimizations: faster load times and improved stability.
Bug Fixes
Fixed a chat synchronization issue along with minor UI corrections.
PWA, Image Generation & Admin Functions
Lurus as Progressive Web App (PWA)
Lurus can now be installed as an app on desktop and mobile – for faster access and a native experience.
AI Image Generation
Create AI-generated images directly in Lurus. Controllable via team policies – owners can enable or disable the feature for the entire team.
Extended Admin Functions
New management features for administrators: improved overview, user management, and extended control options.
Better Chat Titles
Auto-generated chat titles are now more precise and meaningful – for a better overview in chat history.
After-Release Fixes
Bug fixes and stability improvements after the initial launch.
The Launch – Lurus is Live
New Chat Interface
Completely redesigned chat interface with modern design, optimized user experience, and faster loading times.
Microsoft SSO
Single Sign-On with Microsoft accounts – for seamless integration into existing enterprise infrastructure.
New File Icons
Improved visual recognition of file types through new, clear icons in chat.
Voice Input
Speech-to-text function for quick voice input – controllable via team policies.
Settings Redesign
Completely redesigned settings page with clear structure and improved user guidance.
Update to Angular 21
Migration to the latest Angular framework for improved performance and security.
Redesigned Input Fields
All input fields have been visually redesigned for a more consistent and modern look.
Optimized Navigation in Light Mode
Improved color scheme for navigation in light mode for better readability and a more harmonious design.
Want to try Lurus yourself?
Try all new features for free and experience GDPR-compliant AI for your business.
Start for free