Google launches Gemini 4 Neon for fast agentic workflows

Photo: Arwahyn, AI-generated illustration
The Rundown
Google released Gemini 4 Neon today, bringing faster frontier-level reasoning, a 1M context window, and lower costs for complex agentic workflows.
The details
- Google priced Gemini 4 Neon at $3 per million input tokens and $15 per million output tokens.
- Cached input tokens receive a 95% discount at $0.15 per million tokens.
- Gemini 4 Neon features a 1 million-token context window capable of synthesizing PDFs, spreadsheets, audio, and video.
- The model ranked #1 overall on the Artificial Analysis Coding Agent Index with a score of 58.6.
- Google rolled out Gemini 4 Neon across Google AI Studio, the Gemini app, and Google Antigravity for paid tiers.
The full story
Google introduced Gemini 4 Neon today to handle complex software engineering, analytical research, and autonomous agent tasks with higher speed. The model aims to deliver frontier-level intelligence while remaining faster and cheaper to run than massive flagship systems. It features a large 1 million-token context window to process mixed inputs including audio, video, spreadsheets, and dense documents.
Pricing for Gemini 4 Neon is set at $3 per million input tokens and $15 per million output tokens. Cached inputs receive a significant 95% discount, costing $0.15 per million tokens. Google structured the model with four adjustable thinking tiers, allowing teams to balance analytical depth against their budget across varied production workloads.
The model is optimized for persistent, multi-step agent operations. In benchmarks, Gemini 4 Neon ranked first on the Artificial Analysis Coding Agent Index with a score of 58.6 and led on Zapier AutomationBench. It is also designed to act as an orchestrator for parallel subagents inside development environments while maintaining lower fabrication and cheating rates.
Gemini 4 Neon is available immediately in Google AI Studio, Android Studio, the Gemini API, and the Gemini mobile application for paid subscribers. Enterprise access is provided through Gemini for business and Google Antigravity, while specialized cybersecurity features are rolling out to select members of the Fairwind Program.
Why it matters
Gemini 4 Neon signals an industry shift toward practical efficiency, balancing frontier reasoning with lower latency and reduced operational costs. The 95% discount on cached tokens makes long-context agentic operations far more economical for developers. Delivering top benchmark coding scores at a fraction of larger model costs will intensify competition among enterprise AI providers.
What it means for you
Freelancers and agency owners building automated workflows can use Gemini 4 Neon to cut token bills on long-context tasks. The large context window allows you to process entire business ledgers, client reports, or codebases in a single call. You can open Google AI Studio today to test the model's speed on your existing API prompts.
Source: Hacker News, The Decoder
Spot a mistake? Tell us on WhatsApp.