Skip to content
Thursday, 19 November 2026
BREAKING
TECHNOLOGY

Google DeepMind Launches Gemini 4 Neon for Coding and Knowledge Work

Google DeepMind has unveiled Gemini 4 Neon, a coding and reasoning model designed for efficiency in knowledge work and science workflows.

Google DeepMind Launches Gemini 4 Neon for Coding and Knowledge Work
AI-generated illustration

Google DeepMind has released Gemini 4 Neon, a new model aimed at enhancing code generation and knowledge-based tasks. Positioned between Gemini 4 Carbon and Gemini 4 Hydrogen, Neon is designed for efficiency while offering advanced reasoning capabilities.

Neon is intended for knowledge workers, providing near-frontier performance at lower operational costs. The pricing for using Neon is set at $3 per million input tokens and $15 per million output tokens, which Google claims is 32% cheaper per task compared to Carbon and significantly lower than some competitors.

The model can process various input types, including text, images, audio, and video. Although Google has not disclosed its parameter count or specific architecture, it has stated that Neon adapts features from Gemini 4 Carbon.

Neon includes multiple operational levels, with a high setting for general tasks and an xhigh level for more complex workflows, such as engineering and scientific analysis. It supports context compaction for longer runs, allowing for better management of file reads and overall efficiency.

In Google Antigravity, Neon can coordinate multiple Gemini 4 Hydrogen agents, allowing for the division of tasks into smaller units, which improves workflow efficiency. Although Neon performs well on various benchmarks, it faces stiff competition. On the Terminal-bench 4.1 benchmark, Claude Opus 5.5 outperforms Neon with a score of 65.8% compared to Neon’s 65.2%.

On the Vals Index, Neon achieved a score of 70.9%, placing it second in the rankings. Google reports a low misaligned action rate of 0.09% for Neon, indicating a high level of safety in its operational protocols.

Gemini 4 Neon demonstrates Google DeepMind’s commitment to enhancing AI capabilities in various fields, including coding and scientific research. The developers anticipate that users will explore innovative applications using this new model.

Sources