Opera adopts Google Gemini to power its browser AI assistant, image generator.

Opera, the developer of the fifth most popular desktop browser, lets users of its built-in AI assistant, Aria, easily create images using Imagen 2, Google’s latest image generation model.

The Browser Project announced Tuesday a deal to support Aria with Google Gemini for text-based chatbots.

“We are pleased to announce this deepening of our collaboration in the field of generative AI to further strengthen our suite of browser AI services,” Per Wetterdal, Opera’s head of partnerships, said in an official statement. He mentioned years of cooperation. .

The feature is now available immediately in Opera Developer, a publicly accessible version of Opera that allows users to preview and test future features. Developer releases are typically migrated to Opera Beta (formerly Opera Next) before finally becoming the next major stable release.

The update also allows Opera’s AI to speak its responses in a conversational way, thanks to Google’s text-to-audio model.

You can access older versions of Aria in the Opera mobile browser, but the Gemini-based update is not yet available for smartphones. The opera did not respond to a request for comment. Decrypt.

Opera’s AI assistant Aria, first released in May 2023, relied on OpenAI’s ChatGPT but later incorporated a fine-tuned version of Google’s PaLM 2 mode. This model is now outdated as Google transitions its AI development to Gemini, a new baseline model that powers its own AI services. Aria in Opera processes your command and determines which model to use for the task: Gemini or Imagen 2.

By integrating Google’s Gemini model, Opera can now provide users with high-quality responses on par with those achieved with GPT-4. Currently, GPT-4o and Gemini 1.5 Pro are ranked 1st and 2nd on the Chatbot Arena leaderboard, a ranking based on blind user ratings.

“We believe the future of AI is open, and we are providing access to the best of Google infrastructure, AI products, platforms and foundational models to help organizations chart their path through generative AI,” said Eva Fors, Executive Director. .” “We’re excited to strengthen our long-standing collaboration with Opera by powering Opera’s AI innovation within the browser space,” of Google Cloud for the Nordics.

Opera still maintains its existing integration with OpenAI’s ChatGPT (now powered by GPT-4 and Dall-e 3). Users simply need to click on the different icons displayed just below the Aria button.

Opera has been leveraging the potential of browser AI for over a year across all major browsers, including the Opera GX gaming browser. Additionally, the company recently opened a green energy-based AI data cluster in Iceland based on NVIDIA DGX technology to rapidly scale its AI program.

AI Browser Wars

The fight to dominate the browser market has been greatly influenced by the integration of AI. Google’s Chrome is still the clear leader, but its lack of built-in AI integration leaves room for other players to innovate and gain ground.

In particular, Microsoft invested heavily in AI, overhauling its Edge browser, which was once considered malicious by installing many AI-based functions. It was once the subject of ridicule, with users joking that downloading Chrome was the best use for it. Edge has seen a notable resurgence over the past few months.

By switching to a Chromium-based engine and embedding AI features, Microsoft’s browser has climbed the rankings, overtaking Apple’s Safari to take second place by the end of 2023, according to Statcounter data. The turning point coincides with Microsoft’s announcement of a unified Copilot experience on September 26, 2023.

This business transfer between Opera and Google Cloud is important because it provides a Google-based alternative in the AI-powered browser market. With the rise of Microsoft Edge thanks to its AI integration, the collaboration between Opera and Google gives users a clear option to leverage Google’s advanced model instead of OpenAI’s advanced model.

the joy of choice

If the convenience of tapping into an AI tool within your web browser appeals to you, you now have three solid choices. Opera with Aria from Google, Edge with Copilot from Microsoft, Mistral, and Brave with Leo from Anthropic.

Brave, a popular crypto browser, may appeal to users who value privacy. Brave integrates its own AI assistant, Leo, directly into your browser so it can answer questions, provide summaries, generate new content, and more. However, you cannot create an image yet.

Leo is powered by large-scale language models such as Mixtral 8x7B, Claude Instant, and Llama 2 13B. Unlike other AI assistants, Brave hosts these models on its own servers, ensuring that user input and conversations with Leo are not retained or used to train the model.

Between Opera and Edge, Copilot may have the edge for text-based responses, but Google’s Imagen 2 feature trumps Dall-E 3 in terms of realism and consistency. In tests conducted in decryptionAria also proved to be more versatile and creative, understanding shorter prompts.

For example, Aria’s interpretation of the dog’s simple request to eat a hamburger was realistic. Only after Copilot was asked to create a dog wearing sunglasses and eating a hamburger did he ask for a more descriptive prompt, creating a less realistic image with a 3D rendering aesthetic.

If you use Copilot in Edge, it’s worth checking out Opera’s Aria. The image creation functionality of either may be worth trying outside of the private bubble that Leo offers in Brave.

Edited by Ryan Ozawa.