インテリジェンスレポート

Google launches Gemini 4 Neon at $3 per million input tokens for agent work

Google says its new model ranks near the top of independent intelligence and coding indexes while costing less per task than Gemini 4 Carbon.

INSIGHTEX Editorial Desk

2 分で読めます

Google launches Gemini 4 Neon at $3 per million input tokens for agent workインテリジェンスレポート

What happened

Google has released Gemini 4 Neon, a new AI model aimed at software engineering, knowledge work and scientific tasks, Google reported. The company said the model is available from launch in the Gemini API through Google AI Studio and in Gemini for business. Paid Google AI Pro and Ultra subscribers can use it in Google Antigravity, and Pro and Ultra subscribers can use it in the Gemini app.

API pricing is $3 per million input tokens and $15 per million output tokens. Cached input costs $0.15 per million, a 95% reduction. The model has a 1 million-token context window and four thinking levels. According to Google, when set to xhigh, Neon placed third overall on the Artificial Analysis Intelligence Index, earning a score of 61. It also ranked first on the Artificial Analysis Coding Agent Index with a score of 58.6.

Google said that at its default high setting, Neon costs 32% less per task than Gemini 4 Carbon. On some benchmarks, the company said, it costs roughly 50% to 80% less per task than top competing frontier models. A security-focused version, Gemini 4 Neon for cyber, will reach trusted defenders in the coming days through the Fairwind Program, according to the company.

Context

Google uses Gemini 4 Carbon as the reference point for its cost and safety comparisons with Neon. In Antigravity, Neon can act as an orchestrator that directs parallel Hydrogen subagents. Google said Neon retains safeguards equivalent to Carbon's against prompt injection and serious misuse, and that it declined only 1.98% of Terminal-Bench 4.0 tasks.

AI Studio profile

Artificial Analysis profile

出典Google

ブリーフィング

1週間のテックニュースを1通のメールで。

読む価値のある記事を毎週お届けします。配信停止はいつでも可能です。