Google said Neon is designed for complex coding, multistep agentic workflows and analytical reasoning. On the company's own benchmarks, it outscored Carbon on some tasks and beat the midtier Claude Opus 5.5 and GPT-6.2 Sol on tests including knowledge work, automation and cybersecurity.
Artificial Analysis gave Neon a score of 61 on its Intelligence Index, third overall behind Claude Fable 5.2 at 65 and GPT-6.2 Astra at 62. The firm also found a small gain over Carbon at about 17% lower cost per task. Google lists pricing at $3 per million input tokens, $0.15 per million cached input tokens and $15 per million output tokens, which falls between GPT-6.2 Sol and Claude Opus 5.5.
Google AI Pro and Ultra subscribers can use Neon in the Gemini app from today. Developers can reach it through Google's paid API, the Antigravity platform and other business offerings.
Why it matters if you build agents
Agent builders get another midtier option that Google says is faster and cheaper than Carbon, but its per-token price is above GPT-6.2 Sol, so cost per task may matter more than list price. The performance figures come from Google and Artificial Analysis, so test Neon on your own agent workloads before switching models.
Get stories like this every weekday
One short email with the day's agent tooling news.