Google has announced several new versions of its Gemini artificial intelligence models, expanding its AI portfolio with different options for developers and users. The company is making Gemini 2.0 Flash generally available through its AI Studio and Vertex AI platforms, while introducing a new cost-efficient model called Gemini 2.0 Flash-Lite and an experimental version of Gemini 2.0 Pro.
The new Gemini 2.0 Flash Thinking feature, now available in the Gemini mobile app, allows users to interact with Google services like YouTube, Maps, and Search through AI-powered reasoning capabilities. This model breaks down complex problems into smaller steps to provide more accurate results, though it may take longer to generate responses. According to Google, the model can process up to 1 million tokens of context, equivalent to approximately 2,000 pages of text.
Google has also emphasized improvements in safety and security measures for these new models. The company states that Gemini 2.0 uses reinforcement learning techniques to critique its own responses and includes automated security testing to protect against potential vulnerabilities. The new Flash-Lite model matches Gemini 1.5 Flash in speed and price while offering better performance on various benchmarks, with pricing set at $0.075 per million input tokens and $0.30 per million output tokens.
Sources: Google, The Verge, VentureBeat