AI & ComputeLOT 261119 · 3 min read

Google launches Gemini 4 Neon at $3 per million input tokens

Google launched Gemini 4 Neon at $3 per million input tokens, with a 1,048,576-token context window, broad paid-tier availability and vendor-reported benchmark leads.

By Grace Kim645 words

Key takeaways

  1. Gemini 4 Neon launched on the paid Gemini API at $3 per million input tokens and $15 per million output tokens.
  2. Neon provides a 1,048,576-token context window and a default 65,536-token output limit.
  3. The Decoder counted Neon as Google's fourth new Gemini 4 model in four weeks.
  4. Carbon and Argon rates double to $4 input and $20 output per million tokens on Dec. 1.
  5. Google reports scores of 58.6 on Coding Agent Index, 56.3% on AutomationBench and $16,843 on Vending-Bench 2.
Google releases Gemini 4 Neon at $3 per million input tokens, its fourth new model in a month
Google releases Gemini 4 Neon at $3 per million input tokens, its fourth new model in a month — JillWellington / pixabay

By Grace Kim

Google released Gemini 4 Neon on Wednesday at $3 per million input tokens and $15 per million output tokens on the paid Gemini API tier. It joins Carbon, Hydrogen and Argon in Google's paid Gemini API lineup. The model targets complex software engineering, knowledge work and science. The Decoder counted the launch as Google's fourth new Gemini 4 model in four weeks, following Carbon and Hydrogen earlier this month.

Who can access Neon?

Google makes Neon generally available on the paid tier of the Gemini API and also offers the model through developer, enterprise and consumer product surfaces.

  • AI Studio
  • Android Studio
  • Gemini for business
  • Antigravity
  • Gemini app for Google AI Pro and Ultra subscribers

The paid API release is confirmed. The separate cyber version remains a staged rollout for selected defenders.

What does Neon cost?

All figures below are paid-tier prices per million tokens at the standard tier. Google ends Carbon's and Argon's introductory rates on Nov. 30.

Model Input Output Input from Dec. 1 Output from Dec. 1
Gemini 4 Neon $3.00 $15.00 $3.00 $15.00
Gemini 4 Hydrogen $1.50 $7.50 $1.50 $7.50
Gemini 4 Carbon $2.00 $10.00 $4.00 $20.00
Gemini 4 Argon $2.00 $10.00 $4.00 $20.00

Neon keeps its $3 input and $15 output rates after Nov. 30. Hydrogen also holds at $1.50 and $7.50. Carbon and Argon rise from $2 and $10 to $4 and $20 on Dec. 1.

On current rates, Neon sits above Hydrogen on price. It also costs more than Carbon and Argon under their introductory rates, but less than their standard rates beginning Dec. 1.

Neon's cached input costs $0.15 per million tokens. Batch and flex processing cost half the standard rate.

What can Neon handle?

Neon has a context window of 1,048,576 tokens. Output defaults to a maximum of 65,536 tokens. A long decode option lets the model use the full context window for output.

Neon accepts text, images, video, audio and PDFs. It returns text only. It supports code execution, computer use and file search. The model can also ground responses with Google Search and Maps.

How did Neon perform on Google's selected tests?

Google supplied the benchmark results in its release announcement. It reported the following:

  • First on the Artificial Analysis Coding Agent Index, with 58.6.
  • First on Zapier's AutomationBench, with 56.3%.
  • First on Andon Labs' Vending-Bench 2, with a reported result of $16,843.
  • Second on the Vals Index, with 70.9%.

Google also reported a score of 61 on the Artificial Analysis Intelligence Index at the xhigh thinking level. That placed Neon third overall.

At the high thinking level, Google says Neon's cost per task is 32% lower than Carbon's. On safety, Google said Neon “resisted every attack” in Gray Swan's indirect prompt injection benchmark at its highest thinking level, with real-time safeguards enabled.

These are company-reported results. Google cited external benchmark suites, but the release supplied the scores and rankings.

When does the cyber version arrive?

Google will make a separate Gemini 4 Neon for cyber available to vetted defenders in the Fairwind Program in the coming days. Google Cloud account teams will handle access requests.

The program creates a staged, invitation-led release for security specialists. The cyber model stays outside Neon's confirmed general-availability launch across the paid API and listed products.

Neon's commercial appeal will be tested after Nov. 30, when Carbon and Argon double to $4 per million input tokens and $20 per million output tokens while Neon keeps its $3 and $15 rates. Neon will still cost twice Hydrogen's input and output rates, making Google's task-cost claim central to the premium.

mlq.ai (Source)

Filed under

Share this article:

More from Grace Kim

Grace Kim

Show full bio

Market editor covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,138 articles

Read also

« Previous articleNext article »