Summary
On September 2, 2026, Google released Gemini 3.8 Flash, its third Flash model in six weeks, alongside a security-hardened sibling, 3.8 Flash Cyber. Priced identically to 3.7 Flash at 0.75 dollars per million input and 3.75 dollars per million output tokens, it is tuned for long-horizon coding and autonomous agents with a 1M-token context window.
What changed
Google launched Gemini 3.8 Flash with multimodal input across text, image, audio, video, and PDF, a 1M-token context window and 64K output, beating 3.7 Flash on every published benchmark and Claude Opus 5 on three of them. Pricing matches 3.7 Flash but doubles on January 1, 2027. A 3.8 Flash Cyber variant targets security-sensitive deployments, and the model was added to Google's AI Mode and to Vercel's AI Gateway.
Why it matters
A fast, cheap model explicitly built for autonomous agents pressures the price-performance frontier for high-volume agentic workloads, where Flash-class models often run the inner loop. The paired Cyber variant signals that vendors are now productizing security posture as a model SKU, not just a policy.
Evidence excerpt
Google shipped Gemini 3.8 Flash on 2 September 2026. It was the third Flash release in six weeks, alongside a locked-down security sibling called 3.8 Flash Cyber, and it is tuned for long-horizon coding and autonomous agents.