What is Gemini 3.8 Flash?
Gemini 3.8 Flash is Google's Flash-tier model launched on 2026-09-02. It is the third Flash drop in six weeks, priced like 3.7 Flash, with video and audio input and a Fairwind-only Cyber twin.
Gemini 3.8 Flash Explained — What Google Launched on 2026-09-02
TL;DR — On 2026-09-02 Google shipped Gemini 3.8 Flash and 3.8 Flash Cyber, the third Flash drop in six weeks. Intro price stays $0.75 / $3.75 per 1M tokens through 2026-12-31, then $1.50 / $7.50. It takes text, image, video, and speech in. Agentic video understanding is Preview on this SKU — that capability has its own URL. If you just need chapters from a lecture, paste the link; you do not pick a Flash SKU.
GA · 2026-09-02 $0.75 / $3.75 Live demo below
Get the notes, skip the SKU chart
Paste a long-form video URL — BibiGPT turns it into chapters, a transcript, and follow-up Q&A.
Add BibiGPT as a preferred source on Google See more BibiGPT in Top Stories and AI answers. 
1 
2 
3
Key facts (90-second read)
As of 2026-09-05, Google launched Gemini 3.8 Flash and 3.8 Flash Cyber on 2026-09-02. Intro price stays $0.75/$3.75 per 1M tokens through 2026-12-31. It accepts video and audio. Agentic video understanding is Preview on this SKU — that mode already has its own URL. If you need a lecture turned into searchable notes, paste the link into BibiGPT instead of picking gemini-3.8-flash.
Features
What Gemini 3.8 Flash actually is
A new Flash SKU, not a processing flag. Google positioned it as the workhorse for agentic workflows and long-horizon software engineering, with a restricted Cyber twin for trusted defenders.
Third Flash in six weeks (2026-09-02)
Google shipped 3.8 Flash three weeks after 3.7 Flash. Same introductory price as 3.7: $0.75 per 1M input tokens and $3.75 per 1M output tokens through 2026-12-31. Standard rate from 2027-01-01 is $1.50 / $7.50.
Same sticker, more tokens per task
Per-token list price did not move. Independent benches still saw ~30% more output tokens per task and more agentic turns, so cost per task can rise ~40% versus 3.7 Flash. Developers who want to minimize tokens can keep 3.7 Flash.
Video and audio in, text out
Artificial Analysis lists text, image, video, and speech input with text output, and a 1M-token context window unchanged from 3.7 Flash. Official docs also list agentic video understanding as Preview — that mode has its own explainer URL.
What this means if you work with long video
A cheaper, smarter Flash SKU helps teams that already run Gemini themselves. It does not replace a product that files chapters, a transcript, and follow-up Q&A for a human.
You should not have to pick 3.8 vs 3.7
Paste a YouTube or lecture URL, get chapters and a transcript. The point of a video assistant is the artifact — not a model ID on an API dashboard.
DeepSWE is coding, not lecture notes
Google says 3.8 Flash outperforms most larger frontier models on DeepSWE v1.1 long-horizon software engineering. That is an engineering bench. A 90-minute course still needs timestamps and export.
Cyber is not a consumer SKU
3.8 Flash Cyber ships only through the Fairwind Program to trusted government, infrastructure, and software-maintainer defenders. It is not on general Gemini API release. Do not treat it as a public video model.
5 key changes (90-second read)
Headline shifts from Google's 2026-09-02 Flash launch.
- 1
A new SKU, three weeks after 3.7
Gemini 3.8 Flash is the third Flash model in six weeks. Model ID gemini-3.8-flash. Default thinking level is medium. This page is the SKU launch, not the 09-01 agentic-video capability launch.
- 2
Same list price, higher effort can cost more
Intro price matches 3.7 Flash at $0.75 / $3.75 per 1M tokens through 2026-12-31, then $1.50 / $7.50. Google warns the model may use more tokens at higher effort. Independent benches saw ~30% more output tokens per task.
- 3
DeepSWE and Terminal-Bench jumped; HLE did not
Google says 3.8 Flash beats most larger frontier models on DeepSWE v1.1 long-horizon engineering. DataCamp cites Terminal-Bench 2.1 at 90.8% vs 81.6% on 3.7 Flash. Humanity's Last Exam stayed flat (~45%).
- 4
Video, audio, and a Preview mode
Inputs: text, image, video, speech. Output: text. Context: 1M tokens. Official docs list agentic video understanding as Preview. Keep that mode on its 09-01 URL — do not fork a second explainer for the same search.
- 5
Cyber is Fairwind-only
3.8 Flash Cyber is not a public API SKU. Access goes through the Fairwind Program for trusted defenders. Consumer Gemini App, AI Mode, Sheets, AI Studio, and Gemini Enterprise get the standard 3.8 Flash.
3 typical scenarios for BibiGPT users
Where the SKU matters — and where a summarizer is the actual product.
You already run Gemini yourself
Compare 3.8 Flash vs 3.7 Flash on your own agent traces before switching. Then still dump the finished video into BibiGPT so humans get chapters instead of a raw token dump.
You just need the lecture notes
A 90-minute course video does not need a model ID. Paste the URL, export the transcript, ask follow-up questions. That is the BibiGPT path.
You are comparing Omni vs 3.8 vs agentic video
Omni is generate-from-any-input. Agentic video understanding is analyze-the-file-you-already-have. 3.8 Flash is the 09-02 SKU those modes can sit on. Keep each intent on its own URL.
Sources
Launch claims come from Google's own posts plus independent benches.
-
On 2026-09-02 Google launched Gemini 3.8 Flash and 3.8 Flash Cyber at the same introductory price as 3.7 Flash ($0.75 / $3.75 per 1M tokens through 2026-12-31), with DeepSWE long-horizon engineering gains and Fairwind-only Cyber access.
Google — Introducing Gemini 3.8 Flash and 3.8 Flash Cyber ↗ -
Gemini API docs list model ID gemini-3.8-flash, default thinking level medium, and introductory pricing through 2026-12-31; standard $1.50 / $7.50 from 2027-01-01.
Google AI for Developers — Gemini 3.8 Flash generateContent notes ↗ -
Artificial Analysis: 1M context; text/image/video/speech in, text out; ~30% more output tokens per task vs 3.7 Flash despite unchanged per-token pricing.
Artificial Analysis — Gemini 3.8 Flash ↗
What is Gemini 3.8 Flash?
Loved by creators, students & researchers
Why people use BibiGPT to turn videos into text every day.
Trusted by 50,000+ users worldwide
★★★★★
“I paste a link and get clean captions in seconds — it saves me hours of retyping every single week.”
Maya R.
Content Creator · Repurposes short videos
★★★★★
“Exporting the transcript lets me review new words at my own pace instead of pausing the video constantly.”
Daniel K.
Language Learner · Studies with real videos
★★★★★
“Accurate, timestamped text I can quote directly. It has quietly become part of my daily workflow.”
Priya S.
Researcher · Cites public talks
FAQ'S
Frequently Asked Questions
Ask us anything!
Popular guides
1 How to Install Skills in DeepSeek Harness: A Hands-On Guide to Teaching dsh to Watch Videos
Install a video-summary skill into DeepSeek Harness in one copy-paste — SKILL.md matches Claude Code. Or skip dsh: paste a Bilibili or YouTube link in the browser and get a timestamped summary.
2 Bilibili AI Video Summary Tool: BibiGPT Summarizes 30+ Platforms Instantly (2026)
Best Bilibili AI video summary tool 2026? Compare 5 options, then paste a Bilibili link for a free AI recap, mind map, and transcript-style takeaways — no login needed.
3 Bilibili Transcript Tools Compared: Best Subtitle Extractors in 2026
Extract Bilibili subtitles free: paste a BV/av link for a transcript even with no captions. Compare 5 tools for SRT export and AI summary. No signup to start.
Skip the Flash SKU chart. Get the notes.
Paste a long YouTube, lecture, or podcast URL into BibiGPT. You get chapters, a transcript, and follow-up Q&A — without picking Gemini 3.8 Flash, 3.7 Flash, or a Cyber variant.