What happened
Logan Kilpatrick introduces Gemini 4 Neon, saying it pushes the frontier across domains and performs above its model class. He describes it as balancing intelligence, speed, and usability, and says it is available now through Google AI Studio, the API, and more.
Introducing Gemini 4 Neon! Pushing the frontier across domains and punching above its model class in a meaningful way. To me, Neon feels like the right balance of intelligence, speed, and usability. Available now across @GoogleAIStudio, the API, and more!
The story so far
Google launched Gemini 4 Neon, a model optimized for software engineering, knowledge work, and scientific research with a 1-million-token context window and four adjustable thinking levels. Google DeepMind reports that Neon performs 32% more cost-effectively per task than the Carbon model at its default high-thinking setting and ranks first on the Artificial Analysis Coding Agent Index. As of November 18, 2026, the model is available via the Gemini API in Google AI Studio, Gemini for business, and the Gemini app for Google AI Pro and Ultra subscribers; paid Google Antigravity subscribers also have access. Cursor has integrated the model, reporting a 63.7% score on CursorBench 4.1, an increase over the 57.3% scored by Gemini 4 Hydrogen. Pricing is $3 per million input tokens—with a 95% discount to $0.15 for cached input—and $15 per million output tokens.
How the story developed · 3
View full event- 11/19 07:39 (UTC+8)Cursor Makes Gemini 4 Neon AvailableCursor · X
- 11/19 05:56 (UTC+8)Google DeepMind Introduces Gemini 4 Neon for Coding, Analysis, and ScienceGoogle DeepMind
- 11/19 05:37 (UTC+8)
Logan Kilpatrick Introduces Gemini 4 Neon This report
