Google Launches Gemini 4 Neon for Advanced AI Workflows
On November 18, 2026, Google unveiled Gemini 4 Neon, an AI model that promises frontier-level intelligence and fast end-to-end task completion across various domains, including software engineering, knowledge work, and scientific workflows.
Gemini 4 Neon ranks third overall on the Artificial Analysis Intelligence Index with a score of 61 at its highest thinking level, trailing only larger, slower models. It also holds the second position on the Vals Index and excels in steerability, adapting effectively to mid-task user corrections, as indicated by its second-place ranking on the Agent Arena Index. Additionally, Neon has recorded its lowest fabrication rate in internal testing.
The pricing for Gemini 4 Neon is set at $3 per million input tokens and $15 per million output tokens, with a significant discount for cached input tokens. The model features a context window of 1 million tokens.
Capabilities Across Various Workflows
Gemini 4 Neon is designed to handle enterprise tasks that involve diverse inputs such as spreadsheets, PDFs, charts, audio, and video. It synthesizes these inputs into structured deliverables, achieving high rankings on various benchmarks. For instance, it ranks second on the Vals Index and excels in legal and tax research benchmarks.
In terms of financial and analytical tasks, Neon has shown strong performance, ranking first in several benchmarks related to finance, accounting, and complex constraint following. It also performs well in reading professional charts and has achieved competitive scores in various assessments.
Production-Ready Code and Workflow Efficiency
Engineered for long-running agent workflows, Neon maintains performance through complex tasks, whether operating independently or directing subagents in Google Antigravity. It minimizes redundant file reads and efficiently manages context during extended sessions.
Neon ranks first on the Artificial Analysis Coding Agent Index and achieves high scores on various coding benchmarks, demonstrating its capability in multi-file debugging and refactoring tasks.
Cyber Defense and Security Features
With a focus on cybersecurity, Gemini 4 Neon generates secure code and excels in penetration testing, achieving state-of-the-art performance in identifying vulnerabilities and validating risks. It also offers efficient code hardening capabilities, making it a valuable tool for cybersecurity professionals.
Safety and Dependability
Gemini 4 Neon incorporates safety improvements from the Gemini 4 safety and security stack, maintaining protections against misuse while enhancing model alignment and behavior. It has demonstrated resilience against indirect prompt injection attacks and maintains low rates of misalignment and cheating.
Deeper Analysis and Research Capabilities
In the Gemini app, Neon excels at complex reasoning tasks, providing high-quality writing and reliable research capabilities. It efficiently processes dense materials and incorporates current details from various sources, making it suitable for fast-paced environments.
Availability and Next Steps
Gemini 4 Neon is now available for developers through the Gemini API in Google AI Studio and for enterprises via Gemini for business. Consumers who are Google AI Pro or Ultra subscribers can access it through the Gemini app. Additionally, eligible members of the Fairwind Program will receive access in the coming days.
Source: Google DeepMind