Gemini 3.7 Flash Pricing: $0.75/$3.75 per 1M

What Gemini 3.7 Flash Actually Costs: $0.75/$3.75 per 1M Tokens Through Year-End

Gemini 3.7 Flash costs $0.75 per 1M input tokens and $3.75 per 1M output tokens, an introductory rate that runs through December 31, 2026 and represents half of the original Gemini 3.6 Flash cost per million tokens. Google DeepMind and Google AI Studio confirmed the pricing at launch on August 13, 2026, alongside the model card at deepmind.google. The rate is temporary: it is described explicitly as an "introductory price," so the sticker cost you see today is the discounted one, not the long-term standard.

Key Takeaways

• Introductory API price: $0.75/1M input, $3.75/1M output tokens — confirmed by Google's launch blog.
• Discount window: valid through December 31, 2026; it is 50% of the original 3.6 Flash cost.
• Cheaper than rivals: roughly 2.5x under Claude Sonnet 5 ($2.00/$10.00) and GPT-5.6 Terra ($2.00/$12.00) on the model card's own comparison table.
• 2027 pricing is not officially published; community reports point to a doubling to ~$1.50/$7.50, unconfirmed by Google.
• Model ID gemini-3.7-flash, 1,048,576-token context, 65,536 max output tokens.
• Access spans the Gemini API, AI Studio, Antigravity, Gemini Enterprise, and Gemini Spark.

Gemini 3.7 Flash Pricing Breakdown

The table below reflects standard pay-as-you-go token pricing as published on Google's launch blog and the DeepMind model card, as of August 14, 2026.

Item	Price	Source
Input tokens (introductory)	$0.75 / 1M tokens	blog.google
Output tokens (introductory)	$3.75 / 1M tokens	Same
Introductory window	Through Dec 31, 2026	Same
Standard price (post-window)	Not yet officially announced	—
Reported post-2026 rate (unconfirmed)	~$1.50 / $7.50 per 1M	Community reports
Model ID	gemini-3.7-flash	ai.google.dev
Context window	1,048,576 tokens	DeepMind model card
Max output	65,536 tokens	DeepMind model card

Google's own model card lists a caveat: the input and output figures carry an asterisk ($0.75* / $3.75*), matching the "introductory price" language in the blog. Simon Willison flagged the schedule as unusual, noting the price is "scheduled to double in price on December 31, 2026" — an observation, not an official Google figure. Treat the $1.50/$7.50 standard rate as a Reported Standard Rate until Google publishes it.

What That Costs in Practice

Using the confirmed introductory rate of $0.75/1M input and $3.75/1M output, here is what typical workloads run to. Each cell shows the algebra so you can re-run it against your own token counts.

Scenario	Input	Output	Cost calculation	Total
Single chat call	2K tokens	1K tokens	(0.002 × $0.75) + (0.001 × $3.75)	~$0.00525
Long-context doc read	200K tokens	5K tokens	(0.2 × $0.75) + (0.005 × $3.75)	~$0.169
Coding agent task	50K tokens	20K tokens	(0.05 × $0.75) + (0.02 × $3.75)	~$0.1125
1M-token batch (in)	1M tokens	0	1 × $0.75	$0.75

At the published $0.75/1M input rate, a 10K-token prompt costs $0.0075 in input alone. A high-volume agent that stacks many sequential calls — the pattern Rohan Paul described, where "one task can stack many sequential model calls" — benefits directly from the low per-token floor. That cost profile is why several launch-day testers called it a workhorse rather than a peak-reasoning model.

How Gemini 3.7 Flash Pricing Compares

Google's model card publishes a direct price comparison against three competitors. All figures are per 1M tokens and come from the DeepMind card.

Model	Input / 1M	Output / 1M	Intelligence Index
Gemini 3.7 Flash	$0.75*	$3.75*	56
Gemini 3.6 Flash	$0.75*	$3.75*	52
Claude Sonnet 5	$2.00	$10.00	55
GPT-5.6 Terra	$2.00	$12.00	57
Muse Spark 1.2	$1.25	$4.25	57

Gemini 3.7 Flash undercuts Claude Sonnet 5 by roughly 2.7x on input and 2.7x on output while scoring one point higher on the Artificial Analysis Intelligence Index (56 vs 55). Alvaro Cintas summarized the community read: "gemini 3.7 flash is half the price of 3.6 flash and 2-3x cheaper than sonnet 5." For a deeper look at the mid-tier it undercuts, see our Claude Sonnet 5 deep dive. Developers weighing open-weight alternatives at flat rates can compare against our Kimi K3 price guide, since Bindu Reddy pegged 3.7 Flash as "just below Kimi K3 but above Terra" on capability.

Source: @dr_cintas

Free Tier, Limits, and Access

Google has not published a dedicated free-tier price for Gemini 3.7 Flash in the launch bundle, but access is broad. The model is live in the Gemini API, Google AI Studio, Google Antigravity, the Gemini Enterprise Agent Platform, and Gemini Spark. You can test it directly in the AI Studio playground at aistudio.google.com, which historically offers free-of-charge use under rate limits.

On the API side, the paid tier unlocks context caching and a Batch API that Google's pricing docs describe as a 50% cost reduction. Consumption options for 3.7 Flash include Standard, Flex, and Priority PayGo, plus Provisioned Throughput and Batch inference. Gemini Spark, the 24/7 personal agent, now runs on 3.7 Flash and is available to Google AI Pro and Ultra subscribers across 160 countries — a subscription-bundled access path rather than a per-token one.

Some agent platforms layered temporary discounts on top of the introductory rate at launch. Devin (Cognition) offered an extra 50% off through August 27, 2026, per Poonam Soni and Ashutosh Shrivastava. Max Weinbach reported that inside Antigravity "the usage limits are insanely high." If you want to prototype a comparable chat-tier workhorse through an API, kie.ai hosts Gemini 3.6 Flash, the model 3.7 Flash directly succeeds and prices against.

What We Don't Know Yet

Several pricing details remain open as of August 14, 2026:

• The official 2027 standard rate. Google says only "through the end of the year." The widely repeated $1.50/$7.50 figure comes from community reports, not a Google pricing page.
• Whether cached-input and context-caching discounts apply at a specific published rate for this model version.
• Long-context surcharge behavior. Google's Enterprise pricing notes that inputs over 200K tokens can be billed at long-context rates on some Gemini families; the 3.7 Flash-specific figure is not confirmed in the bundle.

These gaps are why this page carries an "as of" date and will be amended when Google publishes the standard schedule.

Frequently Asked Questions

How much does the Gemini 3.7 Flash API cost?

Gemini 3.7 Flash costs $0.75 per 1M input tokens and $3.75 per 1M output tokens as an introductory price through December 31, 2026, per Google's launch announcement.

Is Gemini 3.7 Flash free?

Gemini 3.7 Flash has no dedicated free-tier price published in the launch bundle, but Google AI Studio historically offers free-of-charge access with limited rate limits, and the model is available to test in the AI Studio playground.

Is Gemini 3.7 Flash cheaper than Claude Sonnet 5?

Yes. Gemini 3.7 Flash costs $0.75 input / $3.75 output per 1M tokens versus Claude Sonnet 5's $2.00 input / $10.00 output, making it roughly 2.5x to 2.7x cheaper per token.

What happens to Gemini 3.7 Flash pricing after 2026?

The $0.75/$3.75 rate is introductory and runs through December 31, 2026. Google has not published an official 2027 price, though community reports suggest standard pricing may double to around $1.50/$7.50.

How much cheaper is Gemini 3.7 Flash than Gemini 3.6 Flash?

Gemini 3.7 Flash launched at 50% of the original Gemini 3.6 Flash cost per million tokens, according to Google DeepMind and Google AI Studio.

Where can I use Gemini 3.7 Flash?

Gemini 3.7 Flash is available in the Gemini API, Google AI Studio, Google Antigravity, Gemini Enterprise Agent Platform, and Gemini Spark, per Google's launch and model card.

What is the model ID for Gemini 3.7 Flash?

The model ID is gemini-3.7-flash, as listed in the Gemini API and Gemini Enterprise Agent Platform documentation.

What to Watch Next

Track three signals. First, whether Google publishes an official standard (post-introductory) price before December 31, 2026 — that confirms or refutes the $1.50/$7.50 community estimate. Second, whether the launch-window partner discounts (Devin through Aug 27, and OpenRouter's temporary reduction) expire on schedule or get extended. Third, whether Google ships a dedicated free-tier quota for 3.7 Flash in AI Studio, which would reshape the low-volume cost picture. This page will be amended as each lands.

Building similar high-throughput chat and agent workloads? On kie.ai you can try Gemini 3.8 Flash and Claude Sonnet 5.