- Lumo 2.0 version rebuilds Proton’s privacy assistant with reasoning modes, image generation/recognition, live cited web search, and persistent memory
- The privacy stack combines cryptography and policy: zero-access encryption protects stored chats and images, while inference-time protection builds on Proton’s no-logging, no-training promises that have been delivered in the past.
- Proton’s Lumo 2.0 Lite and Lumo 2.0 Max score 127% and 240% higher than Lumo 1.4 on the Artificial Analytics Intelligence Index, bringing them closer to the next-generation AI models.
Proton has revealed Lumo 2.0, its updated AI alternative to ChatGPT, Claude and Gemini, focusing first and foremost on privacy, a distinctly different approach than most of its competitors.
The new update is not only smarter than its predecessor at what it does, but it also brings a host of new capabilities: reasoning modes, image generation and recognition, live web search with citations, persistent memory, and customizable assistants.
Lumo 2.0 seeks to do all this while leveraging zero-access, no-logs, no-training encryption, a pitch that makes it attractive to privacy-focused consumers, many of whom are already customers of its VPN product line.
Updates, multiple models and faster performance
The biggest update to Lumo 2.0 is that it is now multi-modal, allowing it to collect information and verify a variety of sources without forcing the user to resort to other AI engines for most tasks.
Proton cites 76% faster speed for “everyday queries,” although it admits that complex tasks still require a considerable amount of time.
Users can also take advantage of “Custom Lumos,” or specially designed wizards that retain instructions in memory while maintaining the promise of encryption that Lumo offers, allowing users to avoid starting from scratch every time they have a query to address.
Users can use the fast, general-purpose Lite model for everyday queries and opt for the more complex Max model for demanding work, or use the Quick and Think modes, which offer twice the context window of its predecessor for larger workloads and greater consistency with more complex questions.
Pricing spans a free tier for what Proton calls daily private use, a $12.99 per month Lumo Plus plan with unlimited chats, projects, advanced imaging and access to the most capable models, and a $14.99 per user Lumo Professional tier for teams.
Lumo is also available for enterprise users and offers the same updates mentioned above, making it a significantly more powerful and intelligent AI tool than when we last reviewed it at TechRadar.
It’s important to note that while Lumo 2.0 is a big improvement over its previous version 1.4, it’s not as close to frontier models as Proton might want it to seem: its model scores 51 on the Artificial Analysis Intelligence Index, where current frontier models reach 59 (GPT 5.6 Sol Max) or 60 (Claude Fable 5) versus its own comparisons showing it much closer to more frontier models. old ones, like GPT 5.5 and Claude Opus 4.8.
This isn’t entirely surprising, given that the underlying technology Lumo uses can be found in its privacy policy. Proton claims to use a combination of Qwen 3.5, GLM 5.2, Image-Turbo and FireRed-Image-Edit-1.1, with GLM 5.2 scores roughly identical to the figures it currently cites.
Despite its limitations against new cutting-edge AI models, Lumo 2.0 remains arguably the most privacy-focused AI approach available to most end users today, and comes considerably closer than its predecessor to what has become an increasingly arduous task of late: offering a competitive privacy-focused alternative to the billion-dollar proprietary AI models created by companies like Google, Anthropic, and OpenAI.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds.
