Pick

Introducing Gemini 4 Neon

·2026.11.19 06:00

Key point

The model targets complex software and knowledge workflows with a 1 million token context window and pricing of $3 per million input tokens.

1 / 10

Details

  • Google DeepMind launched Gemini 4 Neon, a model designed to balance frontier-level intelligence with fast end-to-end task completion for software engineering, knowledge work, and scientific workflows.
  • Neon achieves a 61 score on the Artificial Analysis Intelligence Index at the xhigh thinking level, placing it third overall, and ranks second on the Vals Index and Agent Arena Index.
  • The model offers significant cost efficiency, reducing per-task costs by 50-80% compared to top competing frontier models at the high thinking level, and by 36% at the xhigh level.
  • It leads in specialized benchmarks including Legal Research Bench, Tax Agent Bench, Terminal-Bench Science (70.5%), RiemannBench (84.0%), and Zapier AutomationBench (56.3%).

Performance and Benchmarks

  • Ranks first on Surge AI benchmarks for finance, accounting, and spreadsheet modeling, and achieves 73.5% on Chartography for reading dense professional charts.
  • Scores 75.8% on Agents' Last Exam v1.1 and 58.6 on the Artificial Analysis Coding Agent Index, where it ranks first.
  • In code generation, it records 65.2% on Terminal-Bench 4.1 and 46.2% on FrontierCode 2.0 with default high settings.

Cybersecurity and Safety

  • Gemini 4 Neon for cyber achieves state-of-the-art results on WebExploitBench v1.2 (51.26%) and SEC-bench Pro (90.5%), with 50% lower costs than competing large models.
  • Maintains safety protections equivalent to Gemini 4 Carbon, with a 0.0% attack success rate on the Gray Swan IPI benchmark when using real-time safeguards at the highest thinking level.
  • Records the lowest fabrication rate among Gemini 4 models in internal Google Antigravity tests and a 0.09% misaligned action rate in agent safety suites.

Availability and Pricing

  • Available via the Gemini API in Google AI Studio, Gemini for business, and the Gemini app for Google AI Pro and Ultra subscribers.
  • Priced at $3 per million input tokens (with 95% discount for cached inputs at $0.15) and $15 per million output tokens.
  • Features a 1 million token context window and four distinct thinking levels, with the default high level offering a 32% cost reduction compared to Gemini 4 Carbon.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.