Google has released Gemini 3.8 Flash. This model is based on Gemini 3.7 Flash.
Performance improvements have been achieved in software engineering and agentic knowledge workflows. Support for customizable effort levels to control the balance between cost, quality, and latency continues to be provided.
Regarding inputs, the model supports text, images, audio, and video. The context window is up to 1 million tokens. Output is provided in text format, supporting up to 64K tokens.
The knowledge cutoff date is 2026-03. However, depending on the domain, knowledge may be limited to 2025-01.
Gemini 3.8 Flash is provided via API. It is available on platforms including Google AI Studio, the Gemini API, and Google Cloud Platform.
Source: Gemini 3.8 Flash (Hacker News Frontpage, 2026-09-03)