Google has released Gemini 4 Neon, a new midrange AI model it says finishes most tasks about three times faster than its larger Gemini 4 Carbon model at roughly 60% lower cost, SiliconANGLE reported.
The company announced Neon on Wednesday, Nov. 18, on the Google blog.
Who can use it now
For regular users, Neon is showing up in the Gemini app for paying Google AI Pro and Ultra subscribers. Free users weren’t included in the first rollout.
Developers can use it through the Gemini API in Google AI Studio, and businesses can pick it from the model menu in Gemini for business. Google priced it at $3 per million input tokens and $15 per million output tokens, with a 1 million-token context window.
How it compares
On the independent Artificial Analysis Intelligence Index, Neon scored 61 at its highest reasoning setting, third overall behind Anthropic’s Claude Fable 5.2 at 65 and OpenAI’s GPT-6.2 Astra at 62, SiliconANGLE reported.
Google also says Neon ranks No. 1 on Vending-Bench 2, a test of how well an AI can run a simulated business, and costs 32% less per task than Carbon at default settings. Those cost and speed claims come from Google, and independent testers haven’t yet published their own speed comparisons.
Pricing puts Neon between rivals. SiliconANGLE listed OpenAI’s GPT-6.2 Sol at $2 per million input tokens and $10 per million output tokens, and Anthropic’s Claude Opus 5.5 at $4 and $20.
Why a midrange model matters
The biggest AI models get the headlines, but most everyday requests, like summarizing an email thread or drafting a document, don’t need the most powerful option. A faster, cheaper model that scores close to the top means quicker answers in the app and lower bills for companies building on it.
That’s the bet Google is making with Neon. Whether it pays off will depend on how it performs on real work, not just benchmarks, once more people get access.


