Google launches Gemini 3.8 Flash and its Flash Cyber twin

By Carlos Montiel | Enterprise AI Specialist
Leer en español →
Published: 2026-09-07 | By: Carlos Montiel | Reading time: ~5 min

On September 2, Google shipped its third Flash model in six weeks: Gemini 3.8 Flash, alongside a restricted-access sibling, Gemini 3.8 Flash Cyber, purpose-built to find and patch vulnerabilities.

A Flash built for long-running agents

Gemini 3.8 Flash is the general-purpose model of the new generation: it processes text, images, audio, video and PDFs with a 1-million-token context window and up to 64,000 tokens of output. Google tuned it specifically for long-horizon coding tasks and to drive autonomous agents — the fastest-growing use case among its enterprise customers.

According to benchmarks Google published, 3.8 Flash beats its predecessor 3.7 Flash across every reported test, and outperforms Claude Opus 5 on three of them, narrowing the gap even further between "fast and cheap" models and each lab's flagship.

Flash Cyber: the twin only defenders get to see

The more consequential news isn't the standard model but its Cyber variant: a frontier-level performer at vulnerability detection and automatic patch generation, already credited with finding a 13-year-old bug in Chrome's codebase. Unlike Flash, Cyber has no general release: it's only available to "trusted defenders" through Google's new Fairwind Program, a controlled-access scheme designed to keep the same offensive capability out of the wrong hands.

The decision to gate Flash Cyber rather than release it openly confirms a trend already visible with other labs' cybersecurity models: once a model gets good enough at finding vulnerabilities, the lab that built it starts treating it as dual-use and restricts who can call it via API.

Pricing: a grace window through year-end

Gemini 3.8 Flash costs $0.75 per million input tokens and $3.75 per million output tokens through December 31, 2026. Starting January 1, 2027, both prices double, to $1.50 and $7.50 respectively. This is a promotional launch structure, not the permanent steady-state price.

Gemini 3.8 Flash — pricing Input: $0.75 / 1M tokens (through Dec 31, 2026) → $1.50 from Jan 2027 Output: $3.75 / 1M tokens (through Dec 31, 2026) → $7.50 from Jan 2027 Context: 1M tokens | Max output: 64K tokens Flash Cyber: restricted access via Fairwind Program, no general release

What it means for a company buying AI

The release cadence — three Flash versions in six weeks — is itself a signal for architecture teams: any integration built on a specific Google model needs to be designed assuming pricing and capability changes every few weeks, not every year. Pinning a production system to a version name with no migration plan is a source of silent technical debt.

For companies in Guatemala and Latin America, the promotional pricing window through December 2026 is a concrete opportunity to pilot coding or technical-support agents on 3.8 Flash before the per-token cost doubles. And if your company runs critical infrastructure, it's worth checking whether you qualify for Google's Fairwind Program: early access to a model that finds decade-old vulnerabilities is a real defensive advantage, not just a technical curiosity.
Carlos Montiel
Enterprise AI Solutions Architect
LLMs, Agents & Orchestration Specialist
guatemalia.com/#contacto · info@guatemalia.com

Need to implement AI in your company?

Carlos Montiel is an enterprise AI solutions architect. He implements LLMs, Agents, RAG and orchestrators for companies across Guatemala and Latin America. Reach out for a consultation.

Contact Carlos Montiel

info@guatemalia.com