Gemini 3.5 Flash
Stable id gemini-3.5-flash
As of 2026-08-30 the official model page still lists stable id gemini-3.5-flash: 1,048,576 input, 65,536 output, standard $1.50 / $9.00 per million tokens (output includes thinking). Default thinking level is medium. This page does not claim the id is live on QCode.
Takeaways
Standard price / 1M tokens
Official pricing Standard: $1.50 input, $9.00 output including thinking tokens. Batch / Flex are $0.75 / $4.50. This is not the 3.6 / 3.7 intro rate.
Input limit
Official model table Input token limit. What's new also says a 1M context window.
Output limit
Official model table Output token limit. What's new rounds it to 65k. This page uses 65,536 from the model table.
Default thinking
What's new in Gemini 3.5 Flash: default moved from high on 3 Flash Preview to medium. Options: minimal / low / medium / high. Do not send thinking_budget in the same request.
What 3.5 Flash is
A stable Flash SKU on the Gemini API. Official positioning: sustained frontier performance for agents and coding. Inputs: text, image, video, audio, PDF. Output: text. The capability table includes caching, code execution, function calling, Search grounding, structured outputs, URL context, thinking. Computer use is Preview. No audio generation, image generation, or Live API.
What people still search
3.6 / 3.7 Flash are already GA, with intro pricing $0.75 / $3.75 through 2026-12-31. Searches for 3.5 Flash are now “is my old id still the cheap one?” Official Standard is still $1.50 / $9.00 — not cheaper than 3.6 / 3.7 intro. On 2026-08-30 people still mention 3.5 Flash for writing.
Timeline
Model page Latest update: May 2026. Stable id gemini-3.5-flash; preview alias gemini-3-flash-preview is still listed.
What's new FAQ: knowledge cutoff January 2025. Use Search Grounding for newer facts.
This fetch: model page, pricing table, and What's new (footer 2026-08-26) still match those specs. 3.6 / 3.7 intro pricing runs through 2026-12-31.
Confirmed vs watch-outs
Confirmed
Stable id gemini-3.5-flash; 1,048,576 / 65,536; Standard $1.50 / $9.00; default thinking medium; Computer use Preview. Sources: official model page and pricing, fetched 2026-08-30.
Watch-outs
Do not read 3.5 Flash as the cheaper Flash: 3.6 / 3.7 intro is $0.75 input / $3.75 output through 2026-12-31. A row on /models is not a promise it is callable. This page does not say it is already available on QCode or that you can switch on the same endpoint.
3.5 vs newer Flash
Stay on 3.5 Flash
Your workflow is pinned to gemini-3.5-flash, or you are matching old evals and prompts. Price it at $1.50 / $9.00.
Look at 3.6 / 3.7 Flash
You want intro pricing, or Google’s current workhorse Flash. 3.7 intro is $0.75 / $3.75 through 2026-12-31, then $1.50 / $7.50. Follow Google’s pricing table.
How to call it
On Gemini API / AI Studio / Vertex set model=gemini-3.5-flash. Use thinking_level (default medium). Do not send thinking_budget in the same request. Official docs recommend leaving temperature / top_p / top_k at defaults. Treat Computer use as Preview.
On QCode
This page does not claim gemini-3.5-flash is already callable on QCode, and it does not say you can switch on the same endpoint. Whether the id appears in the live catalog is /models plus your own probe. For Google’s own path, use the Gemini API.
FAQ
What is the model id?
gemini-3.5-flash. The official model page marks it Stable. The preview alias gemini-3-flash-preview is still listed.
What does Google charge now?
Standard: $1.50 input / $9.00 output per million tokens, output including thinking. Batch / Flex $0.75 / $4.50. Priority $2.70 / $16.20. Pricing page fetched 2026-08-30.
What is the default thinking level?
medium, changed from high on Gemini 3 Flash Preview. Options: minimal, low, medium, high. Do not combine with thinking_budget.
What is the knowledge cutoff?
Official FAQ: January 2025. Use Search Grounding for later facts.
Can it generate images or audio?
No. Inputs may be image, video, audio, and PDF; output is text. Image generation, audio generation, and Live API are unsupported. Computer use remains Preview.
Can I call it on QCode?
This page makes no availability promise. Use the live /models list and your own test. Do not read a catalog snapshot as “already available.”
Sources
ai.google.dev model page /gemini-api/docs/models/gemini-3.5-flash (Last updated 2026-07-21 UTC); What's new /gemini-api/docs/whats-new-gemini-3.5 (Last updated 2026-08-26 UTC); pricing /gemini-api/docs/pricing Gemini 3.5 Flash section. Fetched 2026-08-30.
Read the official id and price first
3.5 Flash specs follow Google’s docs. QCode’s catalog is the live /models list.
Related
Gemini 3.5 Pro guide
Same-generation Pro page on this site.
Gemini 3.6 Flash guide
The next Flash SKU and intro pricing.
Gemini 3.7 Flash guide
Google’s current workhorse Flash.
Not affiliated with Google / DeepMind. Prices and windows follow ai.google.dev. This page is not a promise that QCode can call this id.