Gemini 2.5: New Models & Capabilities
- The Gemini 2.5 family of hybrid reasoning models, designed for optimal performance, cost, and speed, has reached a new milestone.
- Organizations including Snap and SmartBear, along with developers at Spline and Rooms, have already been utilizing the latest versions of the Gemini 2.5 models in production for several...
- The new Gemini 2.5 Flash-Lite offers higher quality than the 2.0 flash-Lite in coding, math, science, reasoning, and multimodal benchmarks.
The Gemini 2.5 family of hybrid reasoning models marks a significant leap forward! the Pro and Flash models are now generally available,offering developers the speed and cost efficiency needed for production applications. Concurrently, a preview of Gemini 2.5 Flash-Lite, the fastest and most cost-effective model, is being introduced. Early adopters like Snap and SmartBear are already seeing the benefits.Gemini 2.5 Flash-Lite excels in high-volume tasks and includes advanced features like adjustable thinking budgets. This news, brought too you by News Directory 3, provides a thorough overview of these new releases, highlighting enhancements in coding, science, and multimodal benchmarks.Explore and implement the new Gemini 2.5 secondary_keyword models today. Discover what’s next …
Gemini 2.5 Models: Pro, Flash Go Live; Flash-Lite Previewed
Updated June 17, 2025
The Gemini 2.5 family of hybrid reasoning models, designed for optimal performance, cost, and speed, has reached a new milestone. The 2.5 Pro and Flash models are now stable and generally available, allowing developers to confidently build production applications. A preview of Gemini 2.5 Flash-Lite, the most cost-efficient and fastest 2.5 model,is also being introduced.
Organizations including Snap and SmartBear, along with developers at Spline and Rooms, have already been utilizing the latest versions of the Gemini 2.5 models in production for several weeks. Feedback from these users has contributed to the stable releases of 2.5 Flash and Pro.
The new Gemini 2.5 Flash-Lite offers higher quality than the 2.0 flash-Lite in coding, math, science, reasoning, and multimodal benchmarks. It excels in high-volume, latency-sensitive tasks such as translation and classification. It also features lower latency than both 2.0 Flash-Lite and 2.0 Flash across a range of prompts. Gemini 2.5 Flash-Lite includes capabilities such as adjustable thinking budgets, tool connectivity (like Google Search and code execution), multimodal input, and a 1 million-token context length.
What’s next
Developers can begin building with the preview version of Gemini 2.5 Flash-Lite immediately. Further details about the Gemini 2.5 family of models are available in the Gemini technical report.