Google launched Gemini 4 Neon today, making its AI model three times faster and 60% cheaper. CEO Sundar Pichai said the model offers a balance between advanced reasoning, speed, and cost. It replaces the Argon model in the Gemini app and handles complex tasks like reading spreadsheets, videos, and PDFs.
Gemini 4 Neon is said to offer performance on par with or better than that of Gemini 4 Carbon in coding, reasoning, and agentic tasks.
The model was found to be 60% cheaper and three times faster in initial tests.
Google's CEO Sundar Pichai announced the launch of this new AI model via a post on X today.
The introduction of Gemini 4 Neon has led Google to overhaul its Gemini 4 family of models.
The new model sits between Hydrogen and Carbon, offering a balance between advanced reasoning, speed, and cost efficiency.
It also replaces the Argon model in the selection menu of the Gemini app.
Gemini 4 Neon boasts a context of one million tokens and four levels of thinking: low, medium, high, and extra-high.
It can pull information from spreadsheets, PDFs, images/charts, audio/video in the same context window.
Google claims that this model ranks third on the Artificial Intelligence Index with a score of 61 at Extra High thinking level.
Gemini 4 Neon also secured the second spot on the VALS Index with a score of 70.9% and the Agent Arena Index, scoring 1,336 across more than 4,200 real-world agent sessions.
Google further claims that the model ranks first on the Legal Research Bench, Tax Agent Bench, and Zapier's Automation Bench.
It also scored 58.6% in coding evaluations on the Artificial Analysis Coding Agent Index and 65.2% on Terminal-Bench 4.1.
