Gemini 3.8 Flash Arrives With a Cyber Variant and a Price Clock

Image: Gemini 3.8 Flash and its gated Cyber security variant
Google released Gemini 3.8 Flash on September 2, making it the third Flash-tier release in six weeks and taking it to general availability on day one. It targets software engineering, agentic tasks, and multi-step reasoning, while keeping the same 1-million-token context as Gemini 3.7.
It is also a pricing story. Google is putting an introductory API rate in front of developers, then setting a higher rate for January 2027. Alongside the general model, Gemini 3.8 Flash Cyber adds a security-focused variant that finds vulnerabilities and automatically generates patches through the gated Fairwind Program.
A fast Flash cadence
Three Flash releases in six weeks changes how a team should read an announcement. The question is not only whether 3.8 is useful today, but whether a workflow can absorb a new model before the previous one becomes routine. The unchanged 1-million-token context gives teams already using Gemini 3.7 a stable reference point, while the capabilities still need to be checked against the tasks that matter.
Google says Gemini 3.8 Flash is available in the Gemini app for Pro and Ultra users, AI Mode in Search, Sheets, the Gemini API, AI Studio, Antigravity, Android Studio, and Gemini Enterprise. One release therefore spans consumer, developer, and enterprise surfaces rather than remaining an API-only experiment.
Cyber access and a calendar price
Gemini 3.8 Flash Cyber is gated through the Fairwind Program for governments, critical infrastructure, and core technology platforms. The split matters: a security workflow that finds vulnerabilities and generates patches is not simply another public checkbox. Builders can apply the same principle by separating ordinary coding automation from high-impact security work and making extra permissions explicit.
Through December 31, 2026, the introductory API price is $0.75 per 1 million input tokens and $3.75 per 1 million output tokens. On January 1, 2027, those rates become $1.50 and $7.50. Both sides double when the introductory window closes.

Image: Sundar Pichai — Photo: Lukasz Kobus / European Commission, CC BY 4.0, via Wikimedia Commons
Calling the first rate a land-grab price is an interpretation. The concrete fact is simpler: a system launched under the introductory rate must still make sense after January.
Builder tip: budget for January
Put the January rate in your budget from the first prototype. Measure input and output separately, and save the model version, the context your workflow actually consumes, and the task-level result you need. Then a future Flash release can be compared against evidence rather than a headline price.
For BrainMap users, keep the pricing date and a small evaluation set beside the project notes. The durable asset is a record of what the workflow cost and accomplished, not a screenshot of a launch offer.
Sources: 9to5Google, Gigazine, TBreak.
What do you think? Is an introductory price useful momentum, or just a deadline teams should treat as a warning?
Ready to organize your knowledge with AI?
BrainMap automatically classifies your notes, discovers connections, and builds your personal knowledge graph. Free to start — no credit card required.
Start for FreeRelated Articles

Anthropic’s Threat Report Makes ‘Slow Down’ an Industry Question
Anthropic says it disrupted Claude misuse tied to bioweapons research, Russia-linked cyber espionage against Ukraine, and attempts to extract Claude’s capabilities. A missed hacking incident and a call from Dario Amodei are pushing the question of model-development speed into the open.

Siri AI in iOS 27: Apple Turns Personal Context Into an On-Device Assistant
Apple's Siri AI beta brings context from email, messages, calendar, photos, and notes to iOS 27—with onscreen awareness, cross-app actions, and a local-first privacy model.

Claude Fable 5.1: Cheaper Agent Loops, Tiered Safety by Design
Anthropic released Fable 5.1 and Mythos 5.1 as the same model with different safeguard levels, while lower cache-read pricing changes the economics of long-running agents.