MH-3148MarTech & AI
Gemini 4 Neon launches at $3 per million input tokens
Google launched Gemini 4 Neon for AI Pro and AI Ultra users at $3 per 1M input tokens and $15 per 1M output, claiming Carbon-level results with a 1M-token context.
Wire notes
- Gemini 4 Neon costs $3 per 1M input tokens and $15 per 1M output tokens for AI Pro and AI Ultra subscribers.
- Neon has a 1 million-token context window and four thinking levels: Low, Medium, High and Extra High.
- Neon ranked No. 3 on the Artificial Analysis Intelligence Index with a score of 61 at xhigh.
- Neon ranked first on Zapier’s AutomationBench at 56.3% and Andon Labs’ Vending-Bench 2 at $16,843.
- After November 30, Neon will cost 25% less than Argon and Carbon’s new $4 input and $20 output rates.

By James Calloway
Google is launching Gemini 4 Neon at $3 per 1 million input tokens and $15 per 1 million output tokens for AI Pro and AI Ultra subscribers in the Gemini app.
The rollout follows last week’s releases of Carbon and Hydrogen. Google said Neon offers “frontier-level reasoning for complex, high-frequency workflows faster and more cost efficiently.”
Google says Neon has performance comparable to, if not surpassing, Carbon across coding, reasoning and agentic tasks, but runs at Flash speeds and at lower cost.
What changes in Google’s model lineup?
Google positions Neon between Hydrogen and Carbon in the Gemini 4 family. Before the launch, Google’s lineup included:
- 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 replaces Argon, and Google said 3.1 Pro is “finally going away.” For paid users, 3.5 Flash-Lite now appears as Auto in the Gemini app’s model picker.
Google’s picker now offers:
- Auto: “Adapts to your needs”
- Everyday help: 4 Hydrogen
- Deeper analysis: 4 Neon
- Code & complex work: 4 Carbon
Neon carries the same 1 million-token context window as Hydrogen. It offers four thinking levels: Low, Medium, High and Extra High.
Google says Neon handles complex coding, analytical reasoning and production-grade enterprise workflows. Carbon remains the option for the hardest long-horizon projects, Hydrogen for fast everyday execution, and Neon for a balance of depth and efficiency.
Paid users now have a clearer three-model path from everyday help to the toughest projects.
What can Neon do in the Gemini app?
Google says Neon improves three core workflows:
- Higher-quality writing and knowledge work: Neon produces clear, well-structured responses in a natural, direct tone. It reads spreadsheet, PDF and photo uploads more precisely, turns dense material into helpful reports and summaries, and draws relevant details from Gmail and Docs.
- Reliable research and analysis: Neon frequently uses details from web sources to keep answers current on fast-moving subjects. Google cited comparisons of vendor pricing before a major purchase and complex briefs built with the latest market data as examples.
- Faster help for tougher tasks: Neon handles complex, multistep jobs more quickly and efficiently than larger models while stretching usage limits further. Google cited analyzing billing statements for utility spikes and cross-checking multiple receipts and documents to negotiate refunds.
Neon’s practical gain is more work per task, better document handling and access to current web details.
How did Neon perform in benchmarks?
Google reported these launch benchmarks:
- Neon placed No. 3 overall on the Artificial Analysis Intelligence Index, scoring 61 at xhigh. Google said it ranked ahead of comparable models but just behind larger, slower systems.
- Neon ranked No. 2 for steerability on the new Agent Arena Index, which measures how models respond to mid-task corrections. Its global debut on Arena’s new Agent Arena Index also placed it second, with a score of 1336 across 4,200+ real-world agent sessions.
- Neon reached No. 1 on both the Legal Research Bench and Tax Agent Bench.
- On Surge AI, it ranked first on most benchmarks covering finance, accounting, spreadsheet modeling and complex constraint following.
- Neon scored 73.5% on Chartography for reading dense professional charts and posted 75.8% on Agents’ Last Exam v1.1, on par with leading models.
- Neon ranked first overall on Zapier’s AutomationBench with 56.3% and on Andon Labs’ Vending-Bench 2 with $16,843 in its business-running simulation.
- Google called Neon best in class on Terminal-Bench Science at 70.5%, RiemannBench at 84.0% and HLE-Diamond at 71.3%.
- Neon also ranked first overall on the Artificial Analysis Coding Agent Index at 58.6, scored 65.2% on Terminal-Bench 4.1 and reached 46.2% on FrontierCode 2.0 at its default high setting.
The benchmark set supports Google’s claim that Neon can handle demanding agent and coding work at Flash-like speed and cost.
What does the pricing mean for buyers?
Google charges $3 per 1 million input tokens and $15 per 1 million output tokens for Neon. It says that rate is twice Hydrogen’s price.
Argon and Carbon carry promotional prices of $2 per 1 million input tokens and $10 per 1 million output tokens until November 30. After the promotion, Google says they will cost $4 and $20 per 1 million tokens, respectively, leaving Neon 25% cheaper.
Paid AI Pro and AI Ultra users can also access Neon through Google Antigravity.
The pricing advantage becomes clear after November 30, when Neon remains 25% below the new Argon and Carbon rates.
What changes for paid users?
Neon becomes the app’s deeper-analysis option, while Carbon handles the hardest projects and Hydrogen stays on everyday tasks. After November 30, paid users will have a lower-cost middle tier for coding, research and enterprise analysis.
via 9to5google.com (Source)
More from James Calloway
Show full bio
Market editor covering technology and product news in marketing, advertising, marketing technology, media, SEO and search at Marketing Herald.
364 articles