3.7 Flash or 3.6 Flash
Same intro price, different jump
Official Standard intro is $0.75 / $3.75 on both through 31 Dec 2026. 3.7 sells a coding/agent jump, not a discount. QCode does not currently offer the Gemini series.
Updated 2026-10-08
One key across Claude, GPT and Chinese models, billed per token.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard / AMEX), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
Head to head
Standard intro
Both $0.75 / $3.75, then $1.50 / $7.50 on 1 Jan 2027.
DeepSWE v1.1
Google launch post. FrontierCode 1.1 Main 43.6% vs 34.4%. Vendor numbers.
WebDev Arena Elo
Cited by Google. AutomationBench 30.4% vs 17.0%.
Gap
3.7 is the next Flash about three weeks after 3.6.
What we are comparing
Two adjacent Flash SKUs. Context in the same 1,048,576 / 65,536 class. 3.7 is pitched at first-pass coding, long-horizon agents, documents and automation. 3.6 is still a live workhorse, not a retired id.
Do not shop the discount
On launch day Google moved 3.6 onto 3.7’s intro rate. “3.7 is the cheap upgrade” is stale. Independent reruns are thin — treat the table as vendor claims.
Timeline
3.6 Flash ships.
3.7 GA; 3.6 moves to the same intro.
Both Standard lists return to $1.50 / $7.50.
Confirmed vs careful
Confirmed
Both ids are in Google’s docs. Intro Standard price matches. Coding/automation rows are from the 3.7 launch post.
Treat carefully
Aggregator “fastest” lines are not Google’s words. 3.6 is not announced as EOL on the pricing table.
Pick
Stay on 3.6
The workflow is only validated on 3.6.
Move to 3.7
You want the coding/horizon jump and can live with vendor benches.
How to switch
On the Gemini API, swap model=gemini-3.6-flash or gemini-3.7-flash. Do not send thinking=minimal on 3.7 (official: error).
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).
3.7 vs 3.6 FAQ
Which is cheaper?
Official Standard intro is the same.
Is 3.6 going away?
The pricing page still lists gemini-3.6-flash. No EOL date on that table.
Are the benches independent?
They are from Google’s 13 Aug 2026 post. Vendor claims.
Same context?
3.7 docs: 1,048,576 / 65,536. 3.6 is the same 1M-class Flash.
Both 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).
Default which?
New work: try 3.7. Locked 3.6 pipelines: regress first.
Sources
Google Keyword 13 Aug 2026; pricing tables; 3.7 model page.
Test first, then split traffic
Look up both ids in the same catalog. Intro pricing is identical.
Related
Gemini 3.7 Flash guide
Specs and calls.
Gemini 3.6 Flash guide
Previous SKU.
Gemini 3.7 Flash pricing
Intro window and the double.
Not affiliated with Google. Benches from Google’s launch post.