Gemini 3.7 Flash
Deprecated, Google moved to 3.8 Flash
Deprecated: from 2026-10-08 Google replaces gemini-3.7-flash with gemini-3.8-flash, API requests route to 3.8 Flash, and 3.7 Flash is no longer selectable in AI Studio. 3.7 Flash was the next Flash three weeks after 3.6, pitched at coding and agents: 1,048,576 in, 65,536 out, intro $0.75 / $3.75 through 31 Dec 2026.
Updated 2026-10-22
One key across Claude, GPT and Chinese models, billed per token.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
At a glance
8 Oct 2026
Google deprecated gemini-3.7-flash and replaced it with gemini-3.8-flash: no longer selectable in AI Studio, requests route automatically. It shipped on 13 Aug 2026 with the model card, API docs and intro pricing.
3.8 Flash intro price / 1M tokens
Same intro rate for 3.8 Flash and 3.6 Flash: through 31 Dec 2026, then $1.50 / $7.50 from 1 Jan 2027.
DeepSWE v1.1
Google’s table vs 3.6 Flash at 49.0%. FrontierCode 1.1 Main: 43.6% vs 34.4%. Vendor numbers.
Input cap
65,536 out. Thinking levels: low / medium / high. Official docs: minimal returns an error.
What 3.7 Flash is
The Flash SKU in the Gemini 3 line. Native multimodal: text, image, video, audio, PDF in; text out. Google calls it the most intelligent workhorse so far. Supported: caching, code execution, function calling, Search grounding, structured outputs, URL context. Computer use is still Preview. No audio generation, image generation, or Live API.
What people are watching
Two things matter first: the intro price expires at year-end, and the coding jump vs 3.6. Spark (Google AI Pro / Ultra) switches to 3.7 Flash where Spark is offered; the Gemini app still excludes Spark in the EEA, UK, Switzerland and Nigeria. Independent re-runs are thin — treat launch benches as vendor claims.
Timeline
Gemini 3.6 Flash ships. Google later describes 3.7 as three weeks after that.
3.7 Flash GA. Model card, API docs and intro pricing go live together.
3.7 Flash deprecated, requests route to 3.8 Flash. 3.8 Flash intro ends on 1 Jan 2027: $1.50 in / $7.50 out.
Confirmed vs treat carefully
Confirmed
gemini-3.7-flash was deprecated on 8 Oct 2026 and replaced by gemini-3.8-flash; it is no longer selectable in AI Studio and API requests route to 3.8 Flash. Specs unchanged since 13 Aug 2026: 1,048,576 / 65,536, thinking low / medium / high, no minimal. 3.8 Flash intro $0.75 / $3.75 through 31 Dec 2026, then $1.50 in / $7.50 out. All from Google's docs, checked 21 Oct 2026.
Treat carefully
Third-party “fastest of 186 models” lines come from aggregator sites, not Google. 3.5 Pro is still not a stable public flagship — do not read 3.7 Flash as “Pro finally shipped.”
3.8 (was 3.7) or 3.6 Flash
Stay on 3.6
Your bill is already locked, the workflow is only validated on 3.6, or Spark has not rolled out in your region.
Move to 3.8 (was 3.7)
You want the coding / long-horizon jump and you can live with the rate doubling on 1 Jan 2027. Set model=gemini-3.8-flash; requests to 3.7 are only routed there, so keep 3.7 out of new configs.
How to call it
Gemini API / AI Studio: set model=gemini-3.8-flash (requests to gemini-3.7-flash currently route there). Set thinking to low, medium or high. Treat Computer use as Preview — it is not GA.
On QCode
QCode does not currently offer the Gemini series. Please use Claude, GPT or Chinese models instead (e.g. claude-sonnet-5-5, gpt-6.1-sol, glm-5.3). Want to try a newer model such as Claude or GPT for a few days without a monthly plan? Users on a personal plan buy a booster pack when the day's quota is short; without a plan, buy a Pay-as-you-go Pack from $4 for a one-time quota you use within 7 days (top-ups start at $5).
3.7 Flash FAQ
What is the model id?
gemini-3.7-flash. Note: Google deprecated this id on 2026-10-08; it is no longer selectable in AI Studio and API requests route to gemini-3.8-flash.
What does it cost today?
Intro: $0.75 in / $3.75 out per million tokens through 31 Dec 2026. From 1 Jan 2027: $1.50 / $7.50. With 3.7 Flash deprecated, requests are served by 3.8 Flash, which has the same intro rate; Batch is $0.375 in / $1.875 out.
How much better than 3.6 Flash?
Google’s launch post: DeepSWE v1.1 65.3% vs 49.0%, FrontierCode 1.1 Main 43.6% vs 34.4%, WebDev Arena Elo 1588 vs 1538, AutomationBench 30.4% vs 17.0%. Vendor numbers.
Is Gemini 3.7 Flash deprecated, and what replaces it?
Yes, deprecated. On 8 Oct 2026 Google replaced gemini-3.7-flash with gemini-3.8-flash: it is no longer selectable in AI Studio and API requests route to 3.8 Flash. Specs unchanged. 3.8 Flash went GA on 2 Sep 2026.
Can it generate images or speech?
No. It accepts image, video, audio and PDF as input and returns text. Image generation, audio generation and Live API are not supported.
Can I call it on QCode?
No. QCode does not currently offer the Gemini series. Please use Claude, GPT or Chinese models instead (e.g. claude-sonnet-5-5, gpt-6.1-sol, glm-5.3).
Sources
Google Keyword, 13 Aug 2026, “Introducing Gemini 3.7 Flash”; Google changelog, 8 Oct 2026 (gemini-3.7-flash deprecated, replaced by gemini-3.8-flash, requests route automatically); ai.google.dev model pages for Gemini 3.7 Flash (12 Oct 2026 UTC) and Gemini 3.8 Flash (2 Sep 2026 UTC); official price table: intro $0.75 / $3.75 through 31 Dec 2026, Batch $0.375 / $1.875 each.
One key, multiple model families
Claude, GPT and Chinese models are billed per token; see /models for each model's price.
Related
Gemini 3.8 Flash price and availability
Official intro price $0.75/$3.75 through 2026-12-31, then double; the Cyber variant is limited to the Fairwind Program.
Claude Sonnet 5
The usual default coding model in this price band.
2026 model radar
Prices and SKUs in one place.
Crypto payment guide
Nine coins, 56 networks, zero platform fee for AI coding API credits.
Not affiliated with Google or DeepMind. Benchmarks are from Google’s 13 Aug 2026 launch post; independent reruns are separate.