2026-07-21 Changes confirmed high confidence

Google Ships Three Gemini Models July 21: 3.6 Flash GA, 3.5 Flash-Lite, and Government-Gated 3.5 Flash Cyber

With Gemini 3.5 Pro still missing, Google competed on price and efficiency — cutting 3.6 Flash output pricing to $7.50 per million tokens and restricting a new security model to governments and certified partners.

Google shipped three Gemini models on July 21, 2026, according to Gemini API release notes cited by CallMissed: Gemini 3.6 Flash reaching general availability, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber — a security-focused model with no public API.

Context

The trio arrived while Google's flagship Gemini 3.5 Pro — promised at I/O for June — remained unreleased after a third missed target, a delay Bloomberg attributes to coding performance shortfalls. The Flash wave is Google competing on price, speed and segmentation around its missing flagship.

What changed

• Gemini 3.6 Flash (GA): model ID `gemini-3.6-flash`, priced at $1.50 input / $7.50 output per million tokens — output cut from $9 — with 1M-token context, 64K max output, and up to 17% fewer tokens per task than 3.5 Flash per Google's figures. GA across the Gemini API, AI Studio, Android Studio, Antigravity, the Gemini app and enterprise channels; no shutdown date announced for predecessors.
• Gemini 3.5 Flash-Lite: $0.30/$2.50 per million, roughly 350 tokens/second output, rolling into Search and the Gemini app.
• Gemini 3.5 Flash Cyber: a security model available only to governments and certified partners through a CodeMender pilot — no public API or pricing. Google reports it found 55 confirmed V8 (Chrome engine) issues in pilot use, a vendor figure reported by The Prompt Insider.
• Gemini 4 signal: Google DeepMind product lead Logan Kilpatrick confirmed Gemini 4 pre-training has begun, calling it the team's "most ambitious pre-training run yet."

Why it matters

Google extended July's defining pattern — access tiers as product strategy. Flash Cyber joins Anthropic's Mythos 5 (Project Glasswing) in a growing class of capability-gated security models, while the 3.6 Flash price cut and token-efficiency gains pressure Moonshot's Kimi K3 and OpenAI's Terra in the mid-tier. Confirming Gemini 4 while 3.5 Pro is absent reads as deliberate narrative management ahead of Alphabet's earnings.

Details

Pricing and specs are from release-notes-derived coverage; Coursiv's spec sheet matches the headline figures. The 17% token-efficiency claim and the 55-V8-finding claim are Google's own measurements.

Limitations and caveats

Final pricing should be verified against Google's official pricing page, which the research record flags for confirmation. All efficiency and security findings are vendor-reported; no independent benchmark of 3.6 Flash was available at publication.

Sources

• CallMissed — Gemini Flash vs Kimi K3 API pricing (release-notes-derived) (aggregator)
• The Prompt Insider — Google releases Gemini 3.6 Flash, Flash-Lite, Flash Cyber (aggregator)
• Coursiv — Gemini 3.6 Flash specs and pricing (aggregator)

*Update note: This post was last reviewed on 2026-07-22 — one day after launch. Independent evaluations and final list pricing on ai.google.dev are the next verification checkpoint.*

Drafted with AI assistance from source briefs; reviewed for citation completeness and label accuracy.