Gemini 4 Neon: Google's mid-tier AI model claims Carbon quality at 60% cost

Representative image: Pexels / Pixabay
Google released Gemini 4 Neon on 18 November 2026, saying it matches Gemini 4 Carbon in its tests at three times the speed and 60% of the cost.
Gemini 4 Neon: what happened
Google on Wednesday, 18 November 2026 released Gemini 4 Neon, a new artificial intelligence (AI) model that sits between its top-end Gemini 4 Carbon and its fast, low-cost Gemini 4 Hydrogen. In an official blog post by Koray Kavukcuoglu, a senior vice-president at Google DeepMind, the company said that in its early evaluations Neon matched or beat Carbon on overall quality while running about three times as fast at 60% of the cost. The model is rolling out to developers through the Gemini API in Google AI Studio, to businesses through Gemini for business, to paid subscribers in the Gemini app, and in Google's coding tool Antigravity.
Gemini 4 Neon: key facts
- Launch: 18 November 2026, announced on Google's official blog.
- Position in the family: between Gemini 4 Carbon (most capable) and Gemini 4 Hydrogen (fastest everyday model).
- Google's claim: Neon matches or exceeds Carbon on overall quality at about 3x the speed and 60% of the cost, based on Google's own early evaluations.
- Focus areas named by Google: software engineering, knowledge work and science workflows, including multi-file debugging and following multi-step instructions.
- Google says Neon recorded the lowest fabrication rate it has measured in internal testing in Antigravity.
- Availability: Gemini API paid tier via Google AI Studio; Gemini for business; Google AI Pro and Ultra subscribers in the Gemini app and in Antigravity.
Gemini 4 Neon: why it matters
AI companies now sell their models in tiers, much like phones: a flagship that is most capable but slow and expensive, and lighter models that are cheaper and quicker. Neon is Google's attempt to fill the middle with a model that it says performs like the flagship for most everyday work. If that claim holds up in independent testing, developers and companies could move a large share of their AI workload to a cheaper model without losing quality.
For Indian developers and start-ups, cost per task is often the deciding factor. Many Indian AI products, from customer-service bots to coding assistants and document tools, run on paid APIs priced in dollars, so a model that does similar work at a lower price can make a product viable that was too expensive before. Neon's availability in Google AI Studio and the Gemini API means Indian developers can test it now on the paid tier.
Two of Google's claims need careful reading. "Fabrication rate" refers to how often a model invents facts or code that does not exist, sometimes called hallucination. Google says Neon scored the lowest it has recorded, but this comes from Google's internal tests in its own tool, not from independent benchmarks. Likewise, the speed and cost comparison with Carbon is Google's figure. Independent evaluations by developers and research groups over the coming weeks will show how the model performs outside Google's tests.
Peepals calculation: At 60% of Carbon's cost, a workload that cost Rs 1,000 on Carbon would cost about Rs 600 on Neon, a saving of 40%. Combined with roughly three times the speed, Neon would, on Google's figures, do the same work in about a third of the time (1 ÷ 3 = 0.33). Our calculation from the reported figures.
| Gemini 4 model | Role in Google's line-up |
|---|---|
| Carbon | Most capable, top tier |
| Neon (new) | Middle tier; Google says Carbon-level quality at 60% of the cost |
| Hydrogen | Faster, cheaper everyday model |
Gemini 4 Neon: impact
Short-term: Paid Gemini users and developers can switch to Neon now, and companies already paying for Carbon will test whether Neon can handle the same jobs for less.
Long-term: A capable mid-priced model puts pressure on rival AI providers to cut prices for similar performance, which could lower the cost of building AI products in India.
Who is affected: Software developers, AI start-ups, businesses using Gemini for business, and paid Google AI Pro and Ultra subscribers.
Gemini 4 Neon: key takeaway
Developers paying for top-tier AI models should test Gemini 4 Neon on their own tasks before switching, because Google's speed, cost and accuracy claims come from its internal evaluations.
Gemini 4 Neon: questions and answers
What is Gemini 4 Neon?
It is a Google AI model released on 18 November 2026 that sits between Gemini 4 Carbon and Gemini 4 Hydrogen, aimed at coding, knowledge work and science tasks.
Is Gemini 4 Neon free to use?
Google lists it for paid Google AI Pro and Ultra subscribers in the Gemini app and Antigravity, for the paid tier of the Gemini API in Google AI Studio, and for Gemini for business customers.
How does Gemini 4 Neon compare with Gemini 4 Carbon?
Google says that in its early evaluations Neon matched or exceeded Carbon on overall quality at about three times the speed and 60% of the cost. These are Google's own results, not independent tests.
What does "fabrication rate" mean in Google's announcement?
It is how often a model makes up information or code that is not real. Google says Neon had the lowest rate it has recorded in internal testing in its Antigravity coding tool.
Disclaimer: Prepared by the Peepals newsroom from publicly available sources with AI assistance. Information is accurate to the best of our knowledge at the time of publication and may change. Images may be representative. Not professional advice. Report an error via our contact page.
By Anika Sharma, AI & Innovations EditorCovers AI & Innovation and more for Peepals Tech. All stories by Anika Sharma
Your week ahead, every Sunday. Free.
Exam and result dates, IPOs opening, scheme deadlines, government holidays and the RBI decision for the next 7 days, plus the week's top stories. One email a week, unsubscribe in one click.
Anika Sharma is a pen name of the Peepals Global Editorial Team, which researches, fact-checks and publishes every story. The portrait is an illustration.
Responsible for this story: the Peepals Global Editorial Team, led by editor Prashant Sonar. How we report









