Overview
- Google announced Gemini 4 Neon as a specialized member of the Gemini 4 family that is rolling out to developers, enterprises and select paid subscribers through the Gemini app.
- Neon supports a 1,000,000‑token context window and can ingest spreadsheets, PDFs, charts, audio and video so it can consolidate dispersed material into single outputs like reports.
- The model is built to coordinate multi‑agent workflows in Google’s Antigravity environment and act as a scheduler for Hydrogen sub‑agents to run parallel, multi‑step programming tasks such as cross‑file bug finding and refactoring.
- Google published API pricing of $3 per 1M input tokens, $15 per 1M output tokens and $0.15 per 1M cached input tokens, and offered four inference tiers that trade off reasoning, speed and cost.
- Google released vendor benchmarks and internal evaluations claiming Neon matches or exceeds Gemini 4 Carbon on task quality, runs up to three times faster in tests and can lower single‑task cost by about 32%, but independent broad validation of those claims has not been presented.