Google · llm model

Gemini 3.8 Flash

Release Ruler for Gemini 3.8 Flash: released Sep 2, 2026, $0.75 per 1M input tokens, proprietary. Each thin stem on the ruler is another model we track.

Gemini 3.8 Flash is a Google language model released Sep 2, 2026; Google's newest language model in our dataset is Gemini 4 Argon. Proprietary, API access only. Google's current fast Gemini model ("most intelligent Flash model" per Google's changelog), replacing Gemini 3.7 Flash. Verified against vendor sources on Oct 26, 2026.

Reported Last verified · All models · JSON · Markdown

  • ReleasedSep 2, 2026
  • Context window1.05M1,048,576 tokens
  • Max output66K65,536 tokens
  • LicenceProprietaryProprietary (API)
InTextImageVideoAudioPDFOutnot verified

Key facts

Gemini 3.8 Flash: key facts with sources
FactValueSource
VendorGooglen/a
ReleasedSep 2, 2026source
Statusgenerally availablen/a
API idgemini-3.8-flashsource
Context window1,048,576 tokens (1.05M)source
Max output65,536 tokenssource
Inputtext, image, video, audio, pdfsource
LicenceProprietary (API)source
Pricing$0.75 in / $3.75 out per 1M tokens; from Jan 1, 2027: $1.5 / $7.5 Current pricing runs through Dec 31, 2026, then $1.50 / $7.50.source

Where Gemini 3.8 Flash sits in our data

How Gemini 3.8 Flash differs from Gemini 3.5 Flash-Lite

Read from the recorded facts of the two models, not from a head-to-head test. We have not run either.

  • We have a recorded price for Gemini 3.8 Flash but none for Gemini 3.5 Flash-Lite.
  • Both list a 1.05M-token context window.
  • Both accept text, image, video, audio, pdf as input.
  • Gemini 3.8 Flash came out 43 days after Gemini 3.5 Flash-Lite.
  • Facts for the other model: Gemini 3.5 Flash-Lite.

What we tested

Nothing yet. We have not run Gemini 3.8 Flash ourselves, so every fact on this page is Reported from the sources below. When we do test it, the scope will be stated here and in our test logs. See methodology.

Not verified, left out

We could not confirm these on a vendor page, so we do not state them.

  • Output modalities were not captured from the model page.

Sources

More from Google

All 88 models