Google DeepMind launches Gemini 3.8 Flash

· opgehaald 06:14

DeepMind published Gemini 3.8 Flash as generally available—its latest Flash workhorse for agentic coding and knowledge work, building on 3.7 Flash with a 1M context / 64K output window.

On 2 Sep 2026 Google DeepMind listed Gemini 3.8 Flash as generally available on its Flash product page and published a model card the same day. The model is described as the next iteration after Gemini 3.7 Flash, with performance gains for software engineering and agentic knowledge workflows, customizable effort levels, multimodal inputs (text/image/video/audio/PDF), 1M input tokens and 64K output tokens, plus tools such as function calling, search, and computer use. Availability includes the Gemini App, Gemini API, Google AI Studio, Gemini Enterprise Agent Platform, Gemini AI Mode, and Google Antigravity. The same day Google also announced Gemini 3.8 Flash Cyber via its Fairwind Program for trusted cyber defenders (vulnerability detection/patching), alongside the general Flash workhorse at introductory API pricing through end of 2026.