Gemini 3.7 Flash Ships: Google Halves the Price of Its Coding Workhorse
What happened. Google released Gemini 3.7 Flash on August 13, 2026 — just three weeks after Gemini 3.6 Flash — calling it "our most intelligent workhorse model yet for coding and agents." The headline number for anyone who prices AI work: introductory API pricing of $0.75 per 1M input tokens and $3.75 per 1M output tokens through December 31, 2026 — exactly half of Gemini 3.6 Flash's launch price.
The pricing story. From January 1, 2027, pricing steps up to $1.50/$7.50 per 1M tokens — the same rate 3.6 Flash launched at. Context caching runs $0.075 per 1M tokens during the intro period, and the batch API takes 50% off. The free tier stays free. That makes Gemini 3.7 Flash, for the rest of 2026, one of the cheapest capable coding models per token in the market — and a direct input for agency cost estimates. The AI agency pricing calculator now models Gemini 3.7 Flash with this verified pricing, and our model routing comparison carries the updated row.
The benchmark jump that matters
Google reports big gains over 3.6 Flash on exactly the dimensions agencies sell: coding, agentic workflows, and document work.
| Benchmark | Gemini 3.7 Flash | Gemini 3.6 Flash |
|---|---|---|
| FrontierCode 1.1 Main (production code quality) | 43.6% | 34.4% |
| DeepSWE v1.1 (software engineering) | 65.3% | 49.0% |
| WebDev Arena (Arena.ai Elo) | 1588 | 1538 |
| GDP.pdf (document reasoning) | 34.0% | 22.0% |
| AutomationBench (business workflows) | 30.4% | 17.0% |
The AutomationBench jump — 30.4% vs 17.0% — is the one to underline for agency work: it is the benchmark aimed at real-world business workflows, the kind of automation agencies actually build. Google attributes the gains to algorithmic improvements on the 3.6 Flash base, not a new pretraining run.
Specs agencies need for scoping
- 1M-token context window (text, images, audio, video input) — enough for large codebases or long agent sessions.
- 64K-token output (65,536 tokens), including thinking tokens.
- Knowledge cutoff March 2026.
- Access: Gemini API (Google AI Studio, Android Studio), Google Antigravity, Gemini Enterprise Agent Platform, the Gemini Enterprise app, and Gemini Spark for AI Pro/Ultra subscribers in 160+ countries. No open weights.
What to do
Re-baseline every quote that assumes 3.6 Flash pricing. If you are quoting agent builds or coding automation against Gemini Flash, the launch price dropped 50% for the rest of 2026. Run the numbers in the agency pricing calculator with the Gemini 3.7 Flash strategy and check what it does to your margin on fixed-bid work.
Route small tasks to the Flash tier on purpose. A workhorse model that matches the prior tier's price at half the cost is exactly what agent cost blowups are designed to avoid — model choice is a margin lever, and the cheaper-model-class playbook applies here even though Gemini 3.7 Flash is closed-weight.
Note the January cliff in client contracts. The $0.75/$3.75 rate expires December 31, 2026. A quote that assumes intro pricing all year is fine for this year and wrong for next — spell out the $1.50/$7.50 step-up in multi-year retainers, or bake a pricing review into the agreement.
Keep routing discipline. Google itself price-matched 3.6 Flash down to the same intro rate in its pricing docs, so the effective comparison is now 3.7 Flash at 3.6 Flash's old price with better benchmarks — a strict upgrade for the workhorse tier. For the frontier end of the stack, agency cost context and coding agent billing models still anchor the bigger picture.
Frequently asked questions
What is Gemini 3.7 Flash?
Gemini 3.7 Flash is Google's newest workhorse-tier model, released August 13, 2026 — three weeks after Gemini 3.6 Flash. Google describes it as its most intelligent workhorse model yet for coding and agents. It has a 1M-token context window, 64K-token output, and is built on Gemini 3.6 Flash with algorithmic improvements rather than a new pretraining run.
How much does Gemini 3.7 Flash cost?
Introductory pricing through December 31, 2026 is $0.75 per 1M input tokens and $3.75 per 1M output tokens — exactly half of Gemini 3.6 Flash's launch price. From January 1, 2027, pricing rises to $1.50 per 1M input and $7.50 per 1M output. Context caching is $0.075 per 1M tokens and the batch API is 50% cheaper.
How does Gemini 3.7 Flash compare to Gemini 3.6 Flash?
Google reports substantial benchmark gains: FrontierCode 1.1 Main 43.6% vs 34.4%, DeepSWE v1.1 65.3% vs 49.0%, WebDev Arena Elo 1588 vs 1538, GDP.pdf document reasoning 34.0% vs 22.0%, and AutomationBench business workflows 30.4% vs 17.0%.
Should AI agencies build on Gemini 3.7 Flash?
For coding and agent workloads, Gemini 3.7 Flash is now one of the cheapest workhorse options available at $0.75/$3.75 per 1M tokens through 2026. Agencies should re-baseline cost estimates for agent-heavy builds and route small-to-mid tasks to the Flash tier instead of frontier models — but the intro price expires December 31, 2026, so quotes should note the January 2027 doubling.
An agency that routes models for cost — not one that bills by habit
Browse Vetted AI Agencies →Or run Gemini 3.7 Flash through the AI agency pricing calculator first.
Sources
- Google Blog — "Introducing Gemini 3.7 Flash" (Aug 13, 2026): blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/
- Google Gemini API pricing docs (official): ai.google.dev/gemini-api/docs/pricing
- DeepMind model card — Gemini 3.7 Flash: deepmind.google/models/model-cards/gemini-3-7-flash
- 9to5Google — "Gemini 3.7 Flash launches three weeks after last model" (Aug 13, 2026): 9to5google.com/2026/08/13/gemini-3-7-flash-launch/
- Bloomberg — "Google Debuts New Gemini Flash While Top AI Model Still Delayed" (Aug 13, 2026): bloomberg.com/news/articles/2026-08-13/google-debuts-new-gemini-flash-while-top-ai-model-still-delayed
Accuracy note: All pricing and benchmark figures come from Google's official sources (Google blog, Gemini API pricing docs, DeepMind model card) as of Aug 13, 2026, verified against 9to5Google and Bloomberg coverage. "Exactly half of 3.6 Flash's launch price" reflects the $0.75/$3.75 vs $1.50/$7.50 comparison; Google's pricing docs also list 3.6 Flash at the same intro rate through Dec 31, 2026. The intro-price cliff on Jan 1, 2027 and the batch/context-caching rates are from the official pricing docs. Benchmark figures are Google-reported; no independent benchmark verification is claimed. The OpenRouter listing at half the official intro rate ($0.375/$1.875) was noted but not used as canonical pricing. Bloomberg was accessible only in snippet form at writing time; its 3.5 Pro delay angle is not relied on here.