How much does Gemini 3.7 Flash cost?

Introductory pricing is $0.75 per million input tokens and $3.75 per million output tokens, valid through December 31, 2026. From January 1, 2027 the standard rate becomes $1.50/$7.50 per MTok — the same as Gemini 3.6 Flash. The 50% discount is hard-coded to expire, so budget for the step-up.

Is Gemini 3.7 Flash better than Gemini 3.6 Flash for coding?

Yes, substantially. Google reports FrontierCode 1.1 rising from 34.4% to 43.6%, AutomationBench from 17.0% to 30.4%, and GDP.pdf from 22.0% to 34.0% versus 3.6 Flash, with DeepSWE v1.1 at 65.3%. It gains on nearly every published eval while costing half as much during the intro window.

Quick Answer

What Is Gemini 3.7 Flash? Pricing & Specs (Aug 2026)

Q: What is Gemini 3.7 Flash?

Gemini 3.7 Flash is Google's fast, low-cost frontier model released August 13, 2026 — just three weeks after Gemini 3.6 Flash. It has a 1M-token context window, a 65,536-token output limit, and launched at $0.75 input / $3.75 output per million tokens, half the price of its predecessor.

Published: August 16, 2026

The Short Answer

Gemini 3.7 Flash is Google’s fast, cheap frontier model, released August 13, 2026 — only three weeks after Gemini 3.6 Flash (July 21). It ships with a 1M-token context window, a 65,536-token output limit, and introductory pricing of $0.75 / $3.75 per million tokens — exactly half of 3.6 Flash’s $1.50/$7.50. That intro rate runs through December 31, 2026, then reverts to $1.50/$7.50 on January 1, 2027.

Key Facts

	Gemini 3.7 Flash
Vendor	Google DeepMind
Released	August 13, 2026
Intro price (per MTok)	$0.75 in / $3.75 out — through Dec 31, 2026
Standard price (from Jan 1, 2027)	$1.50 in / $7.50 out
Context window	1,000,000 tokens
Max output	65,536 tokens
Predecessor	Gemini 3.6 Flash (July 21, 2026)
Positioning	Coding + agentic workloads at Flash-tier cost

What Improved Over 3.6 Flash

Google’s published deltas are unusually large for a three-week gap:

Benchmark	Gemini 3.6 Flash	Gemini 3.7 Flash
FrontierCode 1.1	34.4%	43.6%
AutomationBench	17.0%	30.4%
GDP.pdf	22.0%	34.0%
DeepSWE v1.1	—	65.3%
Legal Agent Bench (all-pass)	baseline	+2.6 pts

The AutomationBench jump (17.0% → 30.4%) is the one that matters most for agent builders: it measures multi-step task completion, not single-shot code generation. Nearly doubling it is a real capability change, not a benchmark-tuning artifact.

The Pricing Catch

The 50% cut is temporary and dated. Google published an explicit expiry — January 1, 2027 — at which point Gemini 3.7 Flash costs the same $1.50/$7.50 as the model it replaced. If you build cost models on the $0.75/$3.75 number, your unit economics change by 2× on New Year’s Day. Treat the intro window as a migration incentive, not a permanent price.

At intro pricing, a reference 30K-input / 5K-output task costs about $0.041 on 3.7 Flash, versus $0.083 on 3.6 Flash and $0.30 on GPT-5.6 Sol.

Where It Loses

Google’s own comparison table shows GPT-5.6 Terra ($2/$12) edging Gemini 3.7 Flash on several agentic coding evals. Flash-tier models also carry the usual trade-off: excellent throughput and price, but lower reliability than frontier models on long-horizon unattended work where a single failure costs more than the token savings. For that class of job, Claude Opus 5 or GPT-5.6 Sol still win.

Should You Switch?

Already on 3.6 Flash → yes, migrate. Same context, better scores, half the price for the rest of 2026. See the upgrade breakdown.
Cost-optimizing agent loops → benchmark it against GLM-5.3 and GPT-5.6 Terra before committing.
Unattended production refactors → stay on a frontier model; Flash-tier failure rates compound.
Free-tier chat users → this is an API/Gemini-app model, not a self-hosted option.

Last verified: August 16, 2026. Prices from Google’s official pricing announcement.