Continuing to bring you our latest models, with an improved Gemini 2.5 Flash and Flash-Lite release- Google Developers Blog
Sep 25, 2025Shrestha Basu Mallick, Sid Lall, Zach Gleicher, Kate Olszewska

Summary
This post announces updated versions of Gemini 2.5 Flash and 2.5 Flash-Lite, detailing improvements in instruction following, reduced verbosity and costs, stronger multimodal capabilities (audio transcription, image understanding, translation), and enhanced agentic tool use (5% gain on SWE-Bench Verified). It also introduces `-latest` aliases for easier access to the newest models, while maintaining stable versions for production use. The updates are available on Google AI Studio and Vertex AI.
More from the Google Developer Platform trail
- Autonomous LLM post-training with Tunix on TPUs- Google Developers Blog
- HeyGen x Google Cloud: Bringing Avatar IV to TPUs- Google Developers Blog
- Run Ray on TPU, Part 2: Ray AI libraries- Google Developers Blog
- Scaling Agentic RL: High-Throughput Agentic Training with Tunix- Google Developers Blog
- Run Ray on TPU, Part 1: The foundations- Google Developers Blog