Google DeepMind has announced that Gemini 2.5 Flash-Lite, previously available only in preview, is now stable and generally available. The move signals that the model is ready for scaled production use, giving developers a smaller, more economical alternative within the Gemini 2.5 family.
According to DeepMind, the model delivers high quality while remaining cost-efficient and compact. It also inherits key features from the 2.5 generation, including a 1 million-token context window and native multimodality. With this release, teams looking for a lightweight but capable AI model can now deploy it in production without preview-stage caveats.