Google releases Gemini 2.0 for anyone to use, with new low-cost, high-performance and search-enabled versions

Google has announced an update to its AI model family, Gemini 2.0. Gemini 2.0 Flash has been released to the public, Gemini 2.0 Flash Thinking Experimental with apps with search functionality has been released, the Gemini 2.0 Flash-Lite, which is cheaper, is now available in public preview, and a trial version of the high-performance model Gemini 2.0 Pro is now available.
Gemini 2.0 model updates: 2.0 Flash, Flash-Lite, Pro Experimental
https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/
Gemini 2.0: Flash, Flash-Lite and Pro - Google Developers Blog
https://developers.googleblog.com/en/gemini-2-family-expands/?linkId=12807127
Today we're expanding the Gemini 2.0 family with new options and broader availability.
— Google (@Google) February 5, 2025
This builds on the first model we launched in December: 2.0 Flash, our model with low latency and better performance ⚡
Read more on today's launches ⬇ pic.twitter.com/SRpDIJMhUP
The summary of the models that have been released this time is as follows. Gemini 2.0 Flash and Gemini 2.0 Pro Experimental are expected to be able to output images and audio in the future. Gemini 2.0 Flash-Lite can only output text, but it is cost-optimized for use cases that output text on a large scale.

Gemini 2.0 Flash is now publicly available and is available for anyone to use in
the Gemini app . Gemini 2.0 Flash-Lite and Gemini 2.0 Pro are available for trial in Google AI Studio and Vertex AI .

The performance comparison with the Gemini 1.5 family is as follows: The Gemini 2.0 Flash-Lite Preview model has performance that exceeds the Gemini 1.5 Flash. The Gemini 2.0 Flash has evolved to exceed the Gemini 1.5 Pro while maintaining a fast response speed, and the Gemini 2.0 Pro Experimental has further improved performance.

The prices per million tokens for Gemini 2.0 Flash and Gemini 2.0 Flash-Lite are as follows. The previous Gemini 1.5 Flash had a price that varied depending on the length of the prompt, but the 2.0 family has a fixed price regardless of the length of the prompt. Regarding the price guideline, Google states that 'With Gemini 2.0 Flash-Lite, you can generate captions for 40,000 images for less than $1 (about 152 yen).'
Model | Input text, images, and videos | Audio input | Character Output | Context Caching |
---|---|---|---|---|
Gemini 2.0 Flash | 0.1 dollars (approx. 15 yen) | 0.7 dollars (about 106 yen) | 0.4 dollars (about 61 yen) | Non-voice: $0.025 (approx. 4 yen) Voice: $0.175 (about 27 yen) |
Gemini 2.0 Flash-Lite | 0.075 dollars (about 11 yen) | 0.075 dollars (about 11 yen) | 0.3 dollars (about 46 yen) | 0.01875 dollars (approximately 3 yen) |
Gemini 1.5 Flash (Prompt 128,000 or less tokens) | 0.075 dollars (about 11 yen) | 0.075 dollars (about 11 yen) | 0.3 dollars (about 46 yen) | 0.01875 dollars (approximately 3 yen) |
Gemini 1.5 Flash (Prompt over 128,000 tokens) | $0.15 (approx. 23 yen) | $0.15 (approx. 23 yen) | 0.6 dollars (about 91 yen) | 0.0375 dollars (approximately 6 yen) |
At the same time as this announcement, Gemini apps have released '2.0 Flash,' which is no longer a trial version, as well as ' 2.0 Flash Thinking Experimental ,' which strengthens thinking skills through multi-step reasoning, and '2.0 Flash Thinking Experimental with apps,' which adds the ability to access search tools, maps, and information from other sites.

Gemini 2.0 Flash has also been released on GitHub Copilot, allowing even free plan users to try out the improved code generation capabilities.
'We've made exciting advances that improve quality while lowering costs, and Gemini 2.0 is the best model in both cost and performance. This release means that anyone can take advantage of our powerful inference and multimodal capabilities,' said Demis Hassabis, CEO of Google DeepMind. 'This will lay the foundation for our agents.'
Exciting progress as we relentlessly drive quality up & cost down. The Gemini 2.0 series are the leading models for cost & performance. w/today's releases, everyone can tap into their powerful reasoning/multimodal capabilities, which also lay the foundations for our agentic work https://t.co/eRSoedLuTU
— Demis Hassabis (@demishassabis) February 5, 2025
Related Posts:
in Software, Web Service, Web Application, Posted by log1d_ts