Google has officially made its Gemini 2.5 family of artificial intelligence (AI) models generally available, a move that provides users with stable versions of the Gemini 2.5 Pro and Gemini 2.5 Flash models.
Significantly, the Mountain View-based tech giant has extended access to the powerful Gemini 2.5 Pro model to users on the free tier of the Gemini platform. Alongside this, Google has also unveiled Gemini 2.5 Flash-Lite, which the company touts as its fastest and most cost-efficient AI model to date.
The general availability marks a shift from the previous preview status of the Gemini 2.5 Pro and Flash models. During their preview phase, these large language models (LLMs) were prone to errors and glitches, with users unable to fully utilize their capabilities. The stable release is expected to address these issues, offering a more robust and reliable experience.
While Google AI Pro and Ultra users will continue to have access to the Gemini 2.5 Pro model, free tier users can now also utilize it. Google AI Pro users receive an expanded access with a limit of 100 daily prompts, and Ultra users benefit from an even higher rate limit. The company notes that this version of the Pro model is consistent with the one released earlier this month, with no significant changes.
The transition to general availability has also altered the model selector menu on the Gemini website and app. Preview versions of these models are no longer displayed. Free tier users will now see Gemini 2.5 Flash, Gemini 2.5 Pro, and a Personalisation Preview model that can access a user’s Google Search history to provide relevant responses.
The newly introduced Gemini 2.5 Flash-Lite model boasts higher performance than its predecessor, the 2.0 Flash-Lite, demonstrating improved capabilities in areas such as coding, mathematics, science, reasoning, and multimodal tasks. This low-latency model is designed for near real-time applications like translation and classification. It inherits features from the broader 2.5 family, including reasoning across various token budgets, integration with Google Search and code execution tools, multimodal input support, and a substantial context window of one million tokens.
Gemini 2.5 Flash-Lite is currently accessible through Google AI Studio and Vertex AI, which also host the stable versions of Gemini 2.5 Pro and Gemini 2.5 Flash. Furthermore, Google is actively integrating both the 2.5 Flash-Lite and Flash models into its Search functionalities.
