Google has introduced Gemini 4 Neon, a new AI model designed for complex development, research, and data analysis tasks. The model features a 1 million token context window and is capable of processing various file formats, including spreadsheets, PDFs, and multimedia, to generate structured outputs.

In performance benchmarks, Neon achieved a score of 61 in the Artificial Analysis Intelligence Index and secured second place in the Vals Index. It also ranked first in specialized tests such as Legal Research Bench, Tax Agent Bench, and the AutomationBench by Zapier, demonstrating strong capabilities in finance and accounting.

The model is available via Google AI Studio, Gemini API, and Gemini for business. Pricing is set at $3 per million input tokens, $0.15 per million cached input tokens, and $15 per million output tokens. It offers four reasoning levels, allowing users to balance depth and cost.

Neon shows significant efficiency in coding and cybersecurity tasks. It achieved a 58.6 score in the Coding Agent Index and a 51.26% success rate in WebExploitBench v1.2. Compared to the Gemini 4 Carbon model, Neon is reported to be three times faster and costs 60% of the price for similar tasks.