Google today unveiled its latest model, Gemini 2.0, which will form the foundation of agent experiences. The first family member is also known: Gemini 2.0 Flash.

Google launches its latest AI model, Gemini 2.0, which Google says is the most advanced AI model to date with multimodal features such as native image and audio output and integrated use of tools. This should form the basis of agent experiences that can plan, remember and act under your guidance.

The first member of the Gemini 2.0 family is also already known: Gemini 2.0 Flash. This model is available as an experimental model for developers via the Gemini API in Google AI Studio and Vertex AI. Users worldwide can already use the chat-optimized experimental version of 2.0 Flash. The model will be generally available in January in several model sizes.

Gemini 2.0 Flash

Google today introduced its first model within the Gemini 2.0 family: Gemini 2.0 Flash. This model offers low latency and improved performance. Developers can get started with this model immediately via the Gemini API in Google AI Studio and Vertex AI.

In addition, users worldwide can already use a chat-optimized experimental version of 2.0 Flash. You can easily select this from the drop-down menu on desktop and mobile versions.

Gemini 2.0 Flash builds on the success of 1.5 Flash, Google’s current most popular model for developers. According to some key benchmarks, Gemini 2.0 Flash is said to outperform Flash 1.5, including twice the speed.

But this new model has even more to offer. In addition to supporting multimodal inputs such as images, video and audio, Flash 2.0 now also supports multimodal outputs such as directly generated images mixed with text and controllable multilingual text-to-speech (TTS) audio.

“Agency” experience

In addition to launching its latest Gemini model, Google also emphasizes the importance of responsible AI development. “We believe that responsible AI development should occur from the start. “To test how agent experiences can work safely and practically, we are introducing a series of research prototypes and experiments to our trusted community of testers,” Google said.

These research prototypes include Project Astra, a general-purpose AI assistant with multimodal thinking capabilities; Project Mariner, a prototype focused on complex human interactions via Gemini 2.0; Jules, an experimental AI coding agent integrated into GitHub workflows; and domain-specific agents that support both the virtual world of video games and robotics.