Google Unveils Gemini 2.0 Models: A Leap Forward in AI Technology
In December 2023, Google made headlines with the launch of its Gemini 2.0 family of AI models, introducing the highly anticipated Gemini 2.0 Flash. This latest iteration has demonstrated superior performance compared to its predecessor, Gemini 1.5 Pro, particularly in crucial AI benchmarks, while also boasting double the speed.
General Availability of Gemini 2.0 Flash
As of today, developers can start utilizing the Gemini 2.0 Flash model, which now supports advanced features including native tool utilization, a remarkable context window of 1 million tokens, and support for multimodal inputs. The model is capable of generating text, images, and audio outputs. Developers can access this robust model through the Gemini API via Google AI Studio and Vertex AI, facilitating the development of production-level applications. The pricing structure for the Gemini 2.0 Flash model is set at $0.10 per million tokens for text, image, and video inputs and $0.40 for text outputs.
Introducing Gemini 2.0 Flash-Lite
Additionally, Google has announced a new variant of the model known as Gemini 2.0 Flash-Lite. This model retains the same 1 million token context window, positioning itself as the most cost-effective option available. Currently in public preview, the Gemini 2.0 Flash-Lite is priced at $0.075 per million tokens for text, image, and video inputs, with text outputs costing $0.30 per million tokens.
Updates to Gemini 2.0 Pro
Moreover, Google has introduced an experimental update for the Gemini 2.0 Pro model, which features enhanced coding capabilities and improved performance for complex prompts. This advanced variant can manage a 2 million token context window and seamlessly supports tools like Google Search and code execution.
For General Consumers: The Flash Thinking Experimental
On the consumer front, Google has rolled out the Gemini 2.0 Flash Thinking Experimental for users of the Gemini app. Built upon the impressive speed of the Gemini 2.0 Flash model, this version enhances reasoning capabilities, allowing users to follow the model’s logic akin to the recently launched DeepSeek R1 model. This transparency sheds light on the AI’s thought process, enriching the user experience.
Access for Gemini Advanced Subscribers
Subscribers to Gemini Advanced will gain access to the experimental version of the Gemini 2.0 Pro, which is designed to further elevate coding capabilities and tackle more intricate tasks. Both the Gemini 2.0 Flash Thinking Experimental and the 2.0 Pro Experimental models are now progressively being made available to users on the Gemini web and mobile platforms. Google Workspace Business and Enterprise clients can expect access to these innovative models soon.
To explore further details, visit the source for more images and updates.
Leave a Reply ▼