Google DeepMind launches Gemini 4 Neon, priced at $3 per million tokens
Google DeepMind has introduced Gemini 4 Neon, which it says ranks third overall on the Artificial Analysis Intelligence Index with a score of 61. It is rolling out in the Gemini API, Google AI Studio and the Gemini app for paid subscribers.

- Gemini 4 Neon scores 61 on the Artificial Analysis Intelligence Index, ranking third overall.
- Pricing is $3 per million input tokens and $15 per million output tokens.
- It has a 1 million-token context window and four thinking levels.
- It is available in the Gemini API, Google AI Studio, Antigravity and the Gemini app.
Google DeepMind has introduced Gemini 4 Neon, a model it describes as offering frontier-level intelligence with fast end-to-end task completion across software engineering, knowledge work and science workflows.
According to Google DeepMind, Neon ranks third overall with a score of 61 on the Artificial Analysis Intelligence Index at its xhigh thinking level. It ranks second on the Vals Index, where it sits on the Pareto frontier, and second on steerability on the new Agent Arena Index. The company also says Neon recorded the lowest fabrication rate it has yet measured in internal Google Antigravity testing.
On pricing, Google DeepMind lists $3 per million input tokens and $15 per million output tokens. Cached input tokens cost $0.15 per million, a 95% discount. The model has a 1 million-token context window and four thinking levels.
Neon is rolling out in the Gemini API in Google AI Studio and in Gemini for business. Paid Google AI Pro and Ultra subscribers get it in Google Antigravity, and Pro and Ultra subscribers can use it in the Gemini app.
On the Vals Index, Google DeepMind says Neon scores 70.9% and ranks first on the Legal Research Bench and Tax Agent Bench. It reports first place on most Surge AI benchmarks in finance, accounting, spreadsheet modelling and complex constraint following. It also reports 73.5% on Chartography and 75.8% on Agents' Last Exam v1.1.
For coding, the company says Neon ranks first on the Artificial Analysis Coding Agent Index with 58.6. It scores 65.2% on Terminal-Bench 4.1 and 46.2% at its default high setting on FrontierCode 2.0. On SWE-sweep it leads the mini-SWE harness with 12.8% and reaches 39.0% in Google Antigravity.
On cost, Google DeepMind says that at the default high thinking level Neon has a 32% lower cost per task than Carbon. It says cost per task is roughly 50% to 80% lower than top competing frontier models at high thinking, and 36% or more lower at xhigh compared with rivals at their highest setting.
Neon can also act as an Orchestrator directing parallel Hydrogen subagents in Google Antigravity. Demonstrations included a WebGL2 ski simulator and a 100,000-cell Google Sheet built by an agent swarm.
Background
khabr.pk reported on 28 October that Google announced Gemini 4 Carbon, which ranked first on the Artificial Analysis Intelligence Index v5.0 with a score of 59. Carbon began rolling out to the public, as khabr.pk reported on 10 November. On 12 November khabr.pk reported on Gemini 4 Hydrogen, a faster model for coding and agent work.
Why it matters for Pakistan
Pakistani developers and freelancers who use Google's AI tools through the Gemini API or Google AI Studio can now access the new model. Paid Pro and Ultra subscribers in Pakistan can use it in the Gemini app and Antigravity. The listed per-token prices are in US dollars, so costs for Pakistani users depend on the exchange rate, which khabr.pk recorded at Rs276.55 interbank on 18 November. Google says Neon's cost per task is lower than Carbon's at the default setting.
Get khabr.pk news on Facebook and YouTube
Short news videos every day, and the day’s top stories in your feed.



