we_are_coded.by CODE · The world, decoded
БГ

Google · Models

Google releases Gemini 4 Neon: $3 per million input tokens, and by Google's figures the cost per task is 32% below Carbon's

Google released Gemini 4 Neon on 18 November: $3 per million input tokens, $15 per million output tokens and a 1 million-token context window. By Google's figures, at its default thinking level the model costs 32% less per task than Carbon.

Source
Google
Published
The event
Checked on
Responsible editor
Cvetelin Ivanov
How we work
MethodCorrections

Three dollars per million input tokens. That is where the price of Gemini 4 Neon starts. The bill, though, adds up per task. So the most important number in the announcement is a different one: at the default thinking level, high, Neon costs 32% less per task than Carbon. That is what Google says, and it ships on 18 November.

Fact

The facts: on 18 November 2026 Google announced Gemini 4 Neon. The price is $3 per million input tokens ($0.15 per million cached input tokens, a 95% discount) and $15 per million output tokens, with a 1 million-token context window. Rollout begins the same day in the Gemini API in Google AI Studio, in Gemini for business, for paid Google AI Pro and Ultra subscribers in Google Antigravity and for Pro and Ultra subscribers in the Gemini app. By Google's figures Neon ranks #3 overall with a score of 61 on the Artificial Analysis Intelligence Index at its xhigh thinking level, #2 on the Vals Index and #2 globally with 1336 on the new Agent Arena Index. It is #1 on the Artificial Analysis Coding Agent Index with 58.6. At the default high level, by Google's figures, it costs 32% less per task than Carbon. On Vals and Agent Arena, also at high, its cost per task is roughly 50% to 80% lower than that of top competing frontier models.

There are a lot of numbers, and the rankings come from Google itself, restating outside indexes. So keep two. The first is cost per task. The second is the place in Google Antigravity: there Neon can work alone or direct parallel Hydrogen subagents, as Google describes it. One directs, the others execute.

The price per token is not the price of the task.

There is early feedback from a partner inside the announcement itself. In Glean's words, in their first evaluations Neon matched or exceeded Gemini 4 Carbon on overall quality at 3x the speed and 60% of the cost. That is a partner's word, published by Google, not an independent check.

Take a task you already pay Carbon for, run it on Neon too, then put the two bills side by side. Until then, the 32% is a number from the announcement.

The visual is generated code-art, no third-party images.