News

Gemini 4 Neon explained: API pricing, paid access and what the benchmarks mean

Google Gemini 4 Neon launch artwork with an illuminated model name

Google has introduced Gemini 4 Neon, an AI model positioned between its larger Carbon model and faster everyday Hydrogen model. The announcement on November 18 focuses on coding, analysis and long-running work, rather than another free chatbot upgrade. For anyone deciding whether to try it, access and cost matter more than the leaderboard headline.

Who gets Gemini 4 Neon?

Google says Neon is rolling out through the paid tier of the Gemini API in Google AI Studio. Google AI Pro and Ultra subscribers can select it in the Gemini app and try it in Google Antigravity. Google also lists Gemini for business access through its API and model dropdown.

This is a rollout announcement, not a guarantee that every eligible account already has the option. The consumer access described in the launch post is for Pro and Ultra subscribers. It should not be read as a free-tier release. API billing and a consumer subscription are separate access routes, so check the product you plan to use before assuming one payment covers the other.

API pricing: input and output are different bills

The announced API rate is $3 per million input tokens and $15 per million output tokens. Cached input is priced at $0.15 per million tokens. The model has a one-million-token context window.

That means a workload using one million ordinary input tokens and one million output tokens would total $18 at the listed token rates, before any other applicable charges. The cache discount applies to cached input, not every prompt, and does not reduce the listed output rate. These are US-dollar API rates, not an Indian monthly subscription price.

The context window describes how much material a model can work with in a request. It is not a promise that uploading a million tokens is free, that every detail will be recalled perfectly, or that the model can produce a million-token answer.

What is Google claiming?

Google positions Neon for multi-file coding, spreadsheet and document analysis, and work requiring several steps. It says the model can process inputs including PDFs, charts, audio and video, and sustain longer coding workflows. Four thinking levels let developers trade reasoning effort against cost and response time.

At its default high thinking level, Google reports a 32% lower cost per task than Carbon on the Artificial Analysis Intelligence Index comparison in its launch material. It also cites a score of 61 at the higher xhigh thinking level, placing Neon third overall on that index at announcement.

Those are specific benchmark results reported by Google. They are not a universal promise that every job will be faster or cheaper. Changing the thinking level, tools, amount of output or retries can change the real bill. A stronger score at xhigh should not be treated as if it came from the default high setting.

Should you switch a working setup?

For developers, the sensible test is a small set of real tasks: an existing bug, a representative PDF, a spreadsheet you can check, and a task with a known correct answer. Compare completed-task cost and errors, not only response speed. Keep code review, tests and permission checks in place before allowing generated work near a live system.

For Gemini app users, the immediate question is simpler: does the model appear in your paid account, and does it handle your actual work better? This is launch coverage, not a hands-on verdict. Google’s separate Neon-for-cyber offering is restricted to eligible Fairwind Program defenders and should not be confused with general consumer access.

Source: Google’s Gemini 4 Neon announcement. Featured image: Google.

Related posts
News

Siemens Eigen Engineering Agent launches in India: what it does inside TIA Portal

News

Gemini privacy update: keep chat history without AI training, but check retention

News

SiMa.ai Palette 3.0 targets robot AI workloads: what the 3.5x claim means

News

Airtel Travel Week: 10% roaming offer, eligibility and the fine print

Sign up for our Newsletter and
stay informed

Leave a Reply

Your email address will not be published. Required fields are marked *