# Google launches Gemini 4 Neon, taking aim at Claude Opus 5.5 and GPT-6.2 Sol

> **TL;DR:** Neon sits between Gemini 4 Carbon and Gemini 4 Hydrogen, replacing Argon in the Gemini app's model menu for AI Pro and Ultra subscribers The API costs 3 USD per 1 million input tokens and 15 USD per 1 million output tokens, with a 1 million token context window It scores 61 on the Artificial Analysis Intelligence Index, ahead of Opus 5.5 (59) and GPT-6.2 Sol (57), but trails on Terminal-Bench 4.1, FrontierCode 2.0, and Chartography All figures were published by Google and have not been independently verified

- Original: https://migovi.com/en/gemini-4-neon-ra-mat-chi-phi-thap/
- Author: [Saitama](https://migovi.com/en/author/saitama/)
- Published: 2026-11-19T12:56:14+07:00
- Categories: AI, News
- Source: [Digital Trends](https://www.digitaltrends.com/computing/google-says-gemini-4-neon-can-hang-with-claude-opus-5-5-and-gpt-6-2-sol-for-less/)

---

Gemini 4 Neon slots in between Carbon and Hydrogen, matching Carbon's results at 32% lower cost per task, though all figures come from Google itself.

## Where Neon fits in the Gemini 4 lineup

Google has just launched Gemini 4 Neon, a model it says brings frontier-level intelligence to complex work such as coding, knowledge work, and scientific research, without making users wait as long as the larger models do. Google AI Pro and Ultra subscribers can pick Neon right from the model menu in the Gemini app, where it has taken Argon's spot under "Deeper analysis", while Hydrogen sits under "Everyday help".

![Model picker in the Gemini app showing Gemini 4 Neon](https://migovi.com/wp-content/uploads/2026/11/gemini-4-neon-ra-mat-chi-phi-thap-1.jpg)

*Neon appears in the Gemini app's model menu*

Beyond the app, Neon is rolling out gradually on the paid Gemini API in Google AI Studio and Android Studio, for Google Antigravity Pro and Ultra subscribers, and in the enterprise edition of Gemini. For verified enterprise customers, the model is listed in Vertex AI documentation for the global, US, and EU regions.

The naming of this new model family is somewhat confusing, so here is the short version: Neon sits between Gemini 4 Carbon, the model Google reserves for its longest and hardest projects, and Gemini 4 Hydrogen, the faster model for everyday tasks. API pricing is 3 USD per 1 million input tokens (cached input tokens cost 0.15 USD) and 15 USD for output tokens. The context window is 1 million tokens, and there are four reasoning levels, with "high" as the default. Logan Kilpatrick described Neon as a sensible blend of intelligence, speed, and ease of use, adding that it performs well above its weight class.

## How Google compares Neon with its rivals

In a comparison table published by Google itself, Neon scores 61 on the Artificial Analysis Intelligence Index, higher than Claude Opus 5.5 (59) and OpenAI's GPT-6.2 Sol (57). Google says the 61 puts Neon in third place overall, behind larger and slower models. Neon also leads on the Vals Index with 70.9% and on AutomationBench with 56.3%.

![Benchmark table comparing Gemini 4 Neon with Claude Opus 5.5 and GPT-6.2 Sol](https://migovi.com/wp-content/uploads/2026/11/gemini-4-neon-ra-mat-chi-phi-thap-2.jpg)

*Comparison table published by Google*

Neon does not win every test: it trails Opus 5.5 on Terminal-Bench 4.1 and FrontierCode 2.0, and falls behind GPT-6.2 Sol on Chartography. The point Google stresses more is cost. At the default "high" reasoning level, Neon matches Carbon on the same index at 32% lower cost per task, and it is 50 - 80% cheaper per task than the leading frontier models.

![Chart of Gemini 4 Neon's intelligence score against cost per task](https://migovi.com/wp-content/uploads/2026/11/gemini-4-neon-ra-mat-chi-phi-thap-3.jpg)

*Artificial Analysis Intelligence Index score plotted against cost per task*

## What paying Gemini users get

For people who pay for Gemini, Google positions Neon for writing, researching documents, and small everyday chores such as going through a stack of receipts or checking utility bills. The model can handle spreadsheets, PDFs, charts, audio, and video in a single prompt. Keep in mind that these are all Google's own figures and descriptions, so independent tests may well turn up different results.

Neon also connects to Gemini's Tasks, pulling context from Gmail and Docs to do more than heavier models like Carbon can in these kinds of jobs. Pro and Ultra users can also use Neon in Google Antigravity, which has an Orchestrator mode that pairs Neon's planning ability with Hydrogen's execution speed, along with access to Google AI Studio.

**MIGOVI VIEWPOINT**

Neon's biggest appeal is economics rather than winning individual benchmarks. A model that matches Carbon while costing 32% less per task is worth far more to real users, especially those who call the API often. That said, both the score table and the cost figures come from Google, and Neon's losses to Opus 5.5 on hard coding tests show it is not a one-size-fits-all choice. If you are on a Pro or Ultra plan in Vietnam, go ahead and try Neon for everyday work such as writing, summarizing documents, and handling spreadsheets. For heavy coding projects, wait for independent tests before switching entirely.

Source: [Digital Trends](https://www.digitaltrends.com/computing/google-says-gemini-4-neon-can-hang-with-claude-opus-5-5-and-gpt-6-2-sol-for-less/)
