Google launches Gemini 4 Neon, a mid-tier model for coding and high-frequency work

Google on November 18 launched Gemini 4 Neon, the fourth model in its Gemini 4 family, positioned between the fast Hydrogen and the flagship Carbon for complex coding, knowledge work and scientific workflows. API pricing is $3 per million input tokens and $15 per million output tokens, available now across the paid API, the Gemini app and Antigravity.
Google on November 18 launched Gemini 4 Neon, the fourth model in its Gemini 4 family. It is the fourth release in seven weeks: Argon arrived on September 30, Carbon followed about 28 days later, Hydrogen launched on November 12, and Neon comes just six days after Hydrogen. Google positions Neon as a mid-tier model: frontier-level intelligence at a lower cost than its largest models, aimed at complex coding, knowledge work and scientific workflows.
Where Neon sits in the lineup
Under Google's division of labour, Carbon handles the hardest long-horizon reasoning, Hydrogen covers fast everyday execution, and Neon sits in between for complex but high-frequency workflows. Neon offers four thinking levels — low, medium, high and extra high — and Google says that at the default high setting it costs 32% less per task than Carbon. Google also says Neon can handle spreadsheets, PDFs, charts, audio and video in a single context window, and can debug and refactor across multiple files with targeted edits rather than rewriting whole files.
Pricing and availability
Neon is priced at $3 per million input tokens and $15 per million output tokens, with cached input at $0.15 per million — a 95% discount — and a one-million-token context window. It is available now in the paid Gemini API (Google AI Studio and Android Studio), Gemini for business, Antigravity for AI Pro and Ultra subscribers, and the Gemini app for AI Pro and Ultra subscribers, where it takes the "deeper analysis" slot in place of Argon. A cyber-tuned variant will roll out to vetted defenders through Google's Fairwind program.
There is a timing quirk in the pricing: Carbon and Argon are currently on an introductory rate of $2 input and $10 output per million tokens, valid until November 30, doubling to $4 and $20 on December 1. That means Neon costs more per token than Carbon's promotional price until the end of November, and becomes 25% cheaper than Carbon's regular price from December.
Benchmarks, with caveats
On the Artificial Analysis Intelligence Index, Neon scores 61 at its extra-high thinking level, ranking third overall behind GPT-6.2 Astra and Claude Fable 5.2, and one point ahead of Carbon. These benchmark figures are Google's and the testing firm's own claims and have not been independently verified, and Google has not published a breakdown of Neon's scores.
The launch continues Google's recent cadence: a new flagship-class model every few weeks, each cheaper per task than the last. For developers, the real choice is not which model is strongest, but which one makes sense once the price list resets on December 1.
N環球 
💬 Comments
No comments yet. Be the first!