Google launches Gemini 4 Neon with Carbon quality at Flash speeds, live for AI Pro Google launched Gemini 4 Neon on Tuesday, a model the company says matches Gemini 4 Carbon's performance on coding, reasoning and agentic tasks at Flash speeds and lower cost, and made it available in the Gemini app for AI Pro and AI Ultra subscribers. Neon scores 61 on the Artificial Analysis Intelligence Index at its Extra High thinking level (#3 overall), ranks #2 on Arena's new Agent Arena Index with 1336 across 4,200+ real-world agent sessions, and takes #1 on the Legal Research Bench, Tax Agent Bench, Zapier's AutomationBench (56.3%) and Andon Labs' Vending-Bench 2 ($16,843). Neon is priced at $3 per 1M input tokens and $15 per 1M output tokens, twice Hydrogen's price, while Argon and Carbon stay at $2/$10 until November 30 before rising to $4/$20, leaving Neon 25% cheaper; the launch also retires Gemini 3.1 Pro and replaces Argon, with 3.5 Flash-Lite becoming "Auto" in the model picker for paid users. Following last week’s release of Carbon https://9to5google.com/2026/11/10/gemini-app-4-carbon/ and Hydrogen https://9to5google.com/2026/11/12/gemini-4-hydrogen-launch/ , Google today https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/ is rolling out Gemini 4 Neon. Before today, Google’s Gemini 4 family consisted of: - Hydrogen — Everyday help: Flash replacement - Argon — Code & analysis: “…ideal choice for the majority of software engineering and knowledge work” - Carbon — Most complex work: “…built for tasks that test the limits of a model’s intelligence” Neon has performance comparable to Carbon across coding, reasoning, and agentic tasks, but at Flash speeds and a lower cost. As such, Neon is available in the Gemini app for AI Pro and AI Ultra subscribers, with 3.1 Pro finally going away. Neon also replaces Argon. Meanwhile, 3.5 Flash-Lite is now “Auto” in the model picker for paid users. With this launch, the model picker is: - Auto : Adapts to your needs - Everyday help : 4 Hydrogen - Deeper analysis : 4 Neon - Code & complex work : 4 Carbon Neon has the same 1 million-token context window with four thinking levels, like Hydrogen: Low, Medium, High, and Extra High. In terms of benchmarks: - Neon ranks ahead of comparable models and 3 overall with a score of 61 on the Artificial Analysis Intelligence Index https://artificialanalysis.ai/ at its xhigh thinking level, landing just behind larger, slower models. - Neon ranks 2 on steerability adapting cleanly to mid-task user corrections on the new Agent Arena Index https://arena.ai/leaderboard/agent?utm source=social&utm medium=x&utm campaign=lbreboot&utm content=announcement . - …reaches 1 on both the Legal Research Bench https://www.vals.ai/benchmarks/legal research and Tax Agent Bench https://www.vals.ai/benchmarks/tax agent bench . - Neon also shows strong wins across categories on Surge AI https://surgehq.ai/benchmarks , ranking 1 on most benchmarks across finance, accounting, spreadsheet modeling, and complex constraint following. - It also scores 73.5% on Chartography for reading dense professional charts… - …performs on par with leading models on Agents’ Last Exam v1.1 https://agents-last-exam.org/leaderboard 75.8% . - It achieves 1 overall on both Zapier’s AutomationBench https://zapier.com/benchmarks 56.3% and Andon Labs’ Vending-Bench 2 https://andonlabs.com/evals/vending-bench-2 $16,843 , a long-horizon simulation of running a business. - …it debuts 2 globally with a score of 1336 on Arena’s new Agent Arena Index https://arena.ai/leaderboard/agent across 4,200+ real-world agent sessions. - Neon is best in class on Terminal-Bench Science https://www.terminal-bench-science.ai/ 70.5% , RiemannBench https://arxiv.org/abs/2604.06802 84.0% , and HLE-Diamond https://lastexam.ai/blog/hle-diamond 71.3% . - …ranking 1 overall on the Artificial Analysis Coding Agent Index https://artificialanalysis.ai/agents/coding-agents coding-agents-index 58.6 - score of 65.2% on Terminal-Bench 4.1 https://www.tbench.ai/ - 46.2% at its default high setting on FrontierCode 2.0 https://cognition.com/blog/frontier-code-2.0 . Neon is priced at $3/1M input tokens and $15/1M output tokens, or twice as expensive as Hydrogen. Argon and Carbon cost $2/1M input and $10/1M output until November 30. After the promotional pricing, it will be $4/1M input and $20/1M output, with Neon 25% cheaper as a result. FTC: We use income earning auto affiliate links. More. https://9to5mac.com/about/ affiliate our homepage http://9to5google.com/ for all the latest news, and follow 9to5Google on exclusive stories https://9to5google.com/feature/exclusive/ , reviews https://9to5google.com/guides/reviews/ , how-tos https://9to5google.com/guides/how-to/ , and subscribe to our YouTube channel https://www.youtube.com/9to5google