Google DeepMind today announced significant updates to its Gemini 2.0 AI model lineup, introducing new versions that offer enhanced performance as well as models that are accessible to users at no cost.
These developments aim to make Google AI models more widely available in the wake of competitors like OpenAI and DeepSeek offering similar models for free.
Meet Gemini 2.0 Flash-Lite
Addressing the need for affordable AI, Google has introduced Gemini 2.0 Flash-Lite. This model offers better quality than its predecessor, 1.5 Flash, while maintaining the same speed and cost-efficiency. It supports a 1 million token context window and multimodal input, making it suitable for high-volume, high-frequency tasks.
For instance, it can generate relevant one-line captions for approximately 40,000 unique photos at a cost of less than a dollar in Google AI Studio’s paid tier. Gemini 2.0 Flash-Lite is currently available in public preview in Google AI Studio and Vertex AI.
Gemini 2.0 Flash is now generally available
The Gemini 2.0 Flash model is also now generally available through the Gemini API in Google AI Studio and Vertex AI. This model supports multimodal input with text output and features a context window of 1 million tokens, enabling efficient processing of extensive information.
Upcoming enhancements include image generation and text-to-speech capabilities, further broadening its applicability.
Introducing Gemini 2.0 Pro Experimental
Not one to leave out developers, Google has also released an experimental version of Gemini 2.0 Pro. Designed for advanced coding performance and managing complex prompts, this model offers developers an expanded context window of 2 million tokens, allowing for comprehensive analysis of vast information.
Additionally, it can utilize tools like Google Search and code execution to enhance its functionality. Developers can access this experimental model in Google AI Studio and Vertex AI, while Gemini Advanced users can explore it via the Gemini app on both desktop and mobile platforms.
How to access Gemini 2.0
To access the latest models, simply log in to your Gemini AI account and use the drop-down menu in the lefthand corner.
Google DeepMind has always emphasized its dedication to responsible AI development. The Gemini 2.0 lineup incorporates new reinforcement learning techniques, utilizing the model itself to critique its responses, leading to more accurate and targeted feedback. Automated red teaming is also employed to assess safety and security risks, including indirect prompt injection attacks.
These developments highlight Google DeepMind's ongoing efforts to advance AI technology, making it more efficient, accessible, and safe for a broad spectrum of applications.
