Google Ships Three New Gemini Models — and Teases Gemini 4

Google released Gemini 3.6 Flash, 3.5 Flash-Lite, and a gated 3.5 Flash Cyber model on July 21, 2026, while confirming pretraining has begun on Gemini 4.

By Quantara Editorial Team · AI Tools · Published Jul 22, 2026 · 7 min read

Google released three new Gemini models on July 21, 2026: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and a restricted Gemini 3.5 Flash Cyber. Notably absent was Gemini 3.5 Pro, which Bloomberg reported Google has been "struggling to meet internal performance goals" on; the Pro tier has not had a substantive update since February 2026.

Gemini 3.6 Flash

Gemini 3.6 Flash is now the default model in the Flash family. Google prices it at $1.50 per million input tokens and $7.50 per million output tokens, with a one-million-token context window and a 64,000-token output cap. Google reports throughput around 280 tokens per second and says the model uses roughly 17% fewer output tokens than Gemini 3.5 Flash for comparable answers. Its knowledge cutoff is March 2026.

Gemini 3.5 Flash-Lite and Flash Cyber

Gemini 3.5 Flash-Lite is a cheaper, faster tier aimed at high-volume, latency-sensitive use cases. Gemini 3.5 Flash Cyber is a gated cybersecurity-focused model limited to governments and trusted partners in a restricted pilot, rather than general availability.

Where Is Gemini 3.5 Pro?

Google product lead Logan Kilpatrick confirmed the company is testing Gemini 3.5 Pro with partners and hopes it will "land soon," without committing to a date. The gap has left Google's flagship reasoning tier unchanged since February while competitors, including Claude Opus 5 and Kimi K3, shipped major updates in the same week.

Gemini 4 Pretraining Has Started

Kilpatrick also confirmed Google has started its most ambitious pretraining run yet, for Gemini 4, without offering a timeline. That confirmation came alongside the Flash releases rather than as a standalone announcement.

What This Means for Teams on Gemini

Gemini 3.6 Flash's lower output-token usage and steady throughput make it a reasonable default for high-volume production workloads that do not need Pro-tier reasoning depth. Teams that specifically need Gemini's strongest reasoning model should keep evaluating Opus 5, Kimi K3, and GPT-5.6 Sol in parallel until Google ships an updated Pro model.

Build Prompts for Gemini 3.6 Flash

Use Verbito's AI Prompt Generator to write structured prompts that specify task, constraints, and output format, so the same instructions work across Gemini, Claude, and other model families.

Sources

• TechCrunch: Google releases three new Gemini models, but no 3.5 Pro
• 9to5Google: Gemini 3.6 Flash launch
• GitHub Changelog: Gemini 3.6 Flash now available in GitHub Copilot