Gemini 3.7 Flash: Google's Coding Workhorse at Half Price
- Ce qui s’est passé
- Google shipped Gemini 3.7 Flash, its most capable coding and agent workhorse, three weeks after 3.6 Flash.
- Pourquoi c’est important
- It's better on every benchmark Google published and half the price — $0.75/$3.75 intro pricing through 2026.
- Que faire
- Switch from 3.6 Flash before the rate reverts to $1.50/$7.50 on Jan 1, 2027.
Gemini 3.7 Flash is a no-brainer upgrade over 3.6 Flash — better on every benchmark Google published and half the price through 2026. If you're on the Flash line, staying on 3.6 leaves capability and money on the table. (These are Google's own numbers; we haven't independently benchmarked the model.)
What happened
Google shipped Gemini 3.7 Flash three weeks after 3.6 Flash, landing its most capable Flash-tier model yet on coding, document, and automation workloads. The gains are substantive, not cosmetic:
| Benchmark | 3.6 Flash | 3.7 Flash |
|---|---|---|
| FrontierCode 1.1 Main | 34.4% | 43.6% (+9.2 pts) |
| DeepSWE v1.1 | 49.0% | 65.3% (+16.3 pts) |
| WebDev Arena | 1538 Elo | 1588 Elo (+50) |
| GDP.pdf | 22.0% | 34.0% |
| AutomationBench | 17.0% | 30.4% |
Pricing is the second headline: $0.75/$3.75 per 1M tokens (input/output) — exactly half of 3.6 Flash's $1.50/$7.50 — through Dec 31, 2026. On Jan 1, 2027 it reverts to $1.50/$7.50, so the aggressive rate is a launch window, not the permanent list price.
It's live today in Gemini Spark (Google AI Pro/Ultra), the Gemini API via Google AI Studio, GitHub Copilot's model picker, and the Gemini Enterprise Agent Platform. Google also shipped updated CBRN and cyber-offense safeguards on the model.
Why it matters
Google is iterating the Flash line on a three-week cadence and passing the efficiency savings through to price. That's the pace every other frontier lab now has to answer. Gemini 3.6 Flash was already the Flash-tier default; three weeks later the upgrade path is obvious.
What changes for you
- Switch from Gemini 3.6 Flash before Jan 1, 2027, when intro pricing reverts to $1.50/$7.50.
- Try 3.7 Flash in Google AI Studio or GitHub Copilot's model picker.
- Budget for the revert — $0.75/$3.75 is a launch window, not the list price.
FAQ
Is the half price permanent? No. Intro pricing runs through Dec 31, 2026; on Jan 1, 2027 it reverts to $1.50/$7.50.
Where can I use it today? Gemini Spark (Google AI Pro/Ultra), the Gemini API via Google AI Studio, GitHub Copilot's model picker, and the Gemini Enterprise Agent Platform.
Should I switch from 3.6 Flash? Yes — 3.7 Flash is better on every benchmark Google published and half the price through 2026. There's no reason to stay on 3.6 for new work.
Que faire
- 1 Switch from Gemini 3.6 Flash before Jan 1, 2027, when intro pricing reverts to $1.50/$7.50.
- 2 Try 3.7 Flash in Google AI Studio or GitHub Copilot's model picker.
- 3 Budget for the revert — $0.75/$3.75 is a launch window, not the list price.
Outils et modèles concernés
Ne ratez plus jamais une mise à jour
Le récap hebdomadaire — uniquement les changements de verdict et les actions urgentes. Sans remplissage.