Google ships Gemini 4 Neon, a fast reasoning model at $3/$15 per million tokens

The model appeared in Vertex AI's Model Garden with a 1M-token context and four thinking levels; Google says it nears Gemini 4 Carbon quality at Flash-tier speed and cost.
Why it matters
Neon fills the gap between Google's cheap Hydrogen and flagship Carbon models and is the fourth Gemini 4 release in about seven weeks. At $3 per million input tokens and $15 per million output, it costs twice as much as Hydrogen and, until Carbon's introductory pricing ends on 30 November, more than Carbon itself. Google says its quality is close to Carbon on coding, reasoning and agent tasks, but no benchmark scores had been published at launch. Watch for Google's own announcement, independent results such as the Artificial Analysis index, and wider account access.
Key facts
- Model ID: gemini-4-neon
- Price: $3 input / $15 output per million tokens
- 1M-token context; 64K output per segment, up to 1M with Long Decode
- Four thinking levels: Low, Medium, High (default), XHigh
- Appeared in Vertex AI Model Garden around 19:12 UTC on 18 Nov
- Fourth Gemini 4 model after Argon, Carbon and Hydrogen