Google Enhances Gemini 2.5 Models With Deep Think and Native Audio
Google updates its Gemini 2.5 series with Deep Think reasoning for Pro and native audio output for both Pro and Flash. The company also expands developer tools and makes 2.5 Flash widely available.
Google expands its Gemini 2.5 model series with significant new capabilities, making 2.5 Pro the top-ranked model on the WebDev Arena and LMArena leaderboards. The tech giant introduces Deep Think, an experimental enhanced reasoning mode for 2.5 Pro that tackles highly complex math and coding problems. Additionally, both 2.5 Pro and 2.5 Flash now feature native audio output for more natural conversations, advanced security safeguards, and computer use capabilities from Project Mariner.
The company focuses heavily on improving the developer experience by introducing thought summaries in the Gemini API and Vertex AI for increased transparency. Developers gain more control over their projects with extended thinking budgets for 2.5 Pro and new support for MCP tools within the Gemini API and SDK. These tools allow builders to easily access a wider range of open source resources while creating applications.
Google makes Gemini 2.5 Flash available to everyone directly in the Gemini app today. An updated version of 2.5 Flash becomes generally available in Google AI Studio for developers and in Vertex AI for enterprises in early June. The updated 2.5 Pro model follows shortly after, marking another step in Google's ongoing effort to release advanced AI technologies safely and responsibly.