– Opera is the first web browser maker to offer built-in support for 150 local large language models from 50 different model families
– Users of Opera One browser can now swap the proprietary Aria AI service for third-party LLMs and run them locally on their computer for increased privacy and offline access
– Accessing local LLMs within Opera One is currently an experimental feature available through the developer stream of Opera’s AI Feature Drops Program, requiring the latest browser version and 2-10GB of storage for each model
Opera has introduced built-in support for 150 local large language models (LLM) from 50 different model families in its Opera One browser. This feature allows users to run these AI models locally on their computer, ensuring privacy and accessibility without the need for an internet connection. Users can swap Opera’s proprietary Aria AI service for third-party LLMs, offering a wider range of options and control over their data.
This move by Opera outshines other AI-savvy browsers like Microsoft Edge, which only offers AI integrations through Bing and Copilot without local support. Opera’s EVP of browsers and gaming, Krystian Kolondra, sees this as an opportunity for the company to explore the local AI space and build experiences within it. While the feature is currently experimental and only available through the developer stream of Opera’s AI Feature Drops Program, it can be accessed by anyone interested in testing it.
Accessing local LLMs within Opera One is simple, requiring users to navigate to the Aria panel and select a local model from a dropdown menu. The chosen LLM can then be downloaded onto the computer for use in the browser. This feature opens up a range of possibilities for AI enthusiasts and privacy-conscious users who want more control over their AI experiences.