llm-gemini 0.34
Release: llm-gemini 0.34 New model gemini-3.8-flash for Gemini 3.8 Flash , with low, medium and high thinking levels. #146 Fixed async responses failing to record the resolved model version. Thanks, Charlie Tonneslan . #137 Google released Gemini 3.8 Flash (and 3.8 Flash Cyber, but that's available to "trusted defenders" only) today. Here are the pelicans…
Opening of the original on Simon Willison
Summary
Google released Gemini 3.8 Flash, a new model with low, medium, and high thinking levels. This update also fixes an issue with asynchronous responses not recording the resolved model version. The Gemini Flash model is noted for its speed, low cost, and competence in tasks like generating HTML and JavaScript. A new version, 3.8 Flash Cyber, is available to select users. The release includes examples of its capabilities, such as generating HTML code and rendering markdown with SVG.
Why it matters
Why it matters: Google's Gemini 3.8 Flash offers tiered 'thinking levels,' potentially allowing users to balance cost and performance. This release targets developers with its speed and affordability for tasks like code generation. While specific benchmarks against competitors like OpenAI's GPT or Anthropic's Claude are not provided, the emphasis on speed and cost suggests a focus on practical application. The limited release of 'Flash Cyber' hints at future enterprise or specialized applications. Developers using Gemini Flash can expect faster, cheaper code generation and improved handling of specific formats like HTML.
Related: OpenAI: Meta Muse Code & Muse Spark Course – Build AI Agents, APIs, and Full-Stack Apps · Anthropic: Meta Muse Code & Muse Spark Course – Build AI Agents, APIs, and Full-Stack Apps · Microsoft: Gemini 3.8 Flash is now available in GitHub Copilot · Meta: Meta Muse Code & Muse Spark Course – Build AI Agents, APIs, and Full-Stack Apps · xAI: OpenAI Cut Off a Billion-Dollar Customer to Avoid Elon Musk
Rated middle: a real update, not a headline event.
- Google says its new Gemini 3.8 Flash model ‘works harder’ but might cost moreThe Verge AI · Web · Sep 2, 2026
- GOOGLE IS BACK! (Gemini 3.8 Flash)Matthew Berman · Web · Sep 3, 2026
- Google releases Gemini 3.8 Flash, its third Flash model in six weeksArs Technica AI · Web · Sep 2, 2026
- The Most Overhyped and Underhyped New AI ModelsMatt Wolfe · Web · Sep 2, 2026
- The Most Overhyped and Underhyped New AI ModelsMatt Wolfe · Web · Sep 2, 2026
- The Most Overhyped and Underhyped New AI ModelsMatt Wolfe · Web · Sep 2, 2026
- Introducing Gemini 3.8 Flash and 3.8 Flash CyberGoogle DeepMind Blog
- Introducing Gemini 3.8 Flash and 3.8 Flash CyberThe Keyword: Gemini
- Introducing agentic video understanding with GeminiGoogle DeepMind Blog
- Proactive cyber defense for governments and enterprisesGoogle DeepMind Blog
- Gemini 3.8 Flash rolling out three weeks after last release9to5Google
- Agentic video understanding in GeminiGoogle for Developers on YouTube
Questions people ask
- What is new in Gemini 3.8 Flash?
- Gemini 3.8 Flash introduces low, medium, and high thinking levels. It also fixes a bug related to asynchronous response model version recording.
- What are the benefits of Gemini Flash?
- Gemini Flash is fast, cheap, and competent at tasks like generating HTML and JavaScript code.
- Is Gemini 3.8 Flash Cyber available to everyone?
- No, Gemini 3.8 Flash Cyber is currently available only to 'trusted defenders'.
More from Simon Willison 2 more
Page generated Sep 3, 2026. Summaries are Subvolts' own; the story belongs to Simon Willison.