technology
2 min readGoogle introduces Gemini 4 Neon.
Understand what Google announced, who is getting access and what the context-window number actually describes.
Google introduced Gemini 4 Neon on November 18, describing a model for coding, analytical work and enterprise tasks. The company positions it between the depth of Carbon and the everyday speed of Hydrogen. That is the launch's product positioning, not our independent model ranking.
One concrete specification is a one-million-token context window. Tokens are the units a model processes. The window describes how much input and working context can fit, rather than guaranteeing that every detail will be recalled correctly or that a larger upload will produce a reliable answer. Google's launch describes combining documents, spreadsheets, charts, audio and video; we have not independently tested those workflows.
The announced API unit prices are $3 per million input tokens, $0.15 for cached input and $15 for output, in U.S. dollars. These prices belong to the API route. They do not describe what every subscriber can do under a consumer plan or the total cost of completing a project.
Access also has several routes. Google says Neon begins rolling out through the paid Gemini API and specified business/developer tools. It names Google AI Pro and Ultra subscribers for selection in the Gemini app and access in relevant tools. Rollout wording matters: the announcement does not establish that an arbitrary account in an arbitrary country already has the model. The cyber variant has a separate Fairwind eligibility boundary and a rollout over the coming days.
For a reader choosing what to try, the first useful check is the product and account route available to them. Then use a representative task to assess whether the result is useful. Capacity, access and demonstrated quality answer different questions.
