# Gemini 4 Neon

- HackerByte page: <https://hackerbyte.net/en/stories/gemini-4-neon-50552424>
- Original source: <https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/>
- Hacker News discussion: <https://news.ycombinator.com/item?id=50552424>
- Submitted to Hacker News: 2026-11-18T21:34:07.000Z

## AI-generated summary

Google has launched Gemini 4 Neon, a new AI model designed for frontier\-level intelligence and fast end\-to\-end task completion across software engineering, knowledge work, and science. It offers competitive performance, especially in steerability and reduced fabrication, while being more cost\-effective than larger models. Neon is available today through various Google AI platforms and APIs.

## Key points

- Gemini 4 Neon achieves top rankings in AI benchmarks, including the Artificial Analysis Intelligence Index and Agent Arena Index.
- The model boasts a 1 million\-token context window and is priced at $3 per million input tokens and $15 per million output tokens.
- Neon demonstrates strong performance in enterprise, analytical, scientific, and coding workflows, excelling in areas like chart analysis and automated business simulation.
- For cybersecurity, Gemini 4 Neon shows state\-of\-the\-art capabilities in generating secure code, penetration testing, and vulnerability patching.
- It offers four thinking levels for scalable control over intelligence and token spend, providing high performance at a lower cost per task compared to previous models.
