Google's agent-optimized Gemini 4 Neon model is now generally available in GitHub Copilot.
Google has launched Gemini 4 Neon, a new model specifically engineered for agentic coding, multi-step tool utilization, and complex knowledge work, within GitHub Copilot. Designed to balance performance and efficiency, Neon offers quality comparable to the flagship Gemini 4 Carbon model but at a significantly lower token cost and with faster response times. The model is available across all major Copilot surfaces, including VS Code, JetBrains IDEs, and the CLI, with administrative controls for enterprise users to manage access.
This launch underscores Google's strategic pivot toward high-efficiency, task-specific models for the developer experience. By positioning Neon as a lower-cost, high-velocity alternative to Carbon, they are directly addressing the cost-sensitivity of large-scale agentic workflows.
* Reduces token overhead for routine coding tasks.
* Enhances agentic reliability with faster latency.
* Demonstrates seamless integration of proprietary models into third-party IDE ecosystems.
DISCOVERED
16m ago
2026-11-18
PUBLISHED
27m ago
2026-11-18
RELEVANCE
AUTHOR
github