Google released Gemini 3.7 Flash on August 13, 2026, just three weeks after Gemini 3.6 Flash. Google’s own model page describes it as “our most capable Flash model for agentic workflows and multimodal reasoning.” The pricing has a wrinkle worth understanding before you build on it: the attractive rate you’ll see quoted everywhere is introductory, with a confirmed expiry date.
What is Gemini 3.7 Flash?
Gemini 3.7 Flash (gemini-3.7-flash in the API) is the latest entry in Google’s Flash tier — the fast, cost-efficient line sitting below the Pro-tier models. Google’s rapid Flash release cadence has been a defining feature of the Gemini 3.x line: 3.5 Flash launched at Google I/O in May, 3.6 Flash followed in July, and 3.7 Flash arrived in mid-August. Each release has targeted the same use case — high-volume agentic and multimodal workloads where per-token cost matters as much as raw capability.
Availability and API access
Gemini 3.7 Flash is available now through the Gemini API, Google AI Studio, and Vertex AI, using the model string gemini-3.7-flash. It’s also referenced in Google’s Antigravity agent platform documentation as one of the models available for agentic workflows.
Pricing: the introductory rate and its expiry
This is the detail most coverage glosses over. Per Google’s own Gemini API pricing page:
| Tier | Input per 1M tokens | Output per 1M tokens |
|---|---|---|
| Standard, through Dec 31, 2026 | $0.75 | $3.75 |
| Standard, from Jan 1, 2027 | $1.50 | $7.50 |
| Batch, through Dec 31, 2026 | $0.375 | $1.875 |
| Batch, from Jan 1, 2027 | $0.75 | $3.75 |
The $0.75/$3.75 rate you’ll see quoted as “Gemini 3.7 Flash pricing” is explicitly time-limited on Google’s own pricing page — it is not a permanent rate, unlike some other vendors’ “introductory” pricing that later becomes standard (see our note on Claude Sonnet 5’s cancelled price increase for a case where that happened the other way). If you’re budgeting a project that runs past December 31, 2026, plan for the rate to double, not stay flat.
The batch API rate stacks with the introductory pricing: $0.375/$1.875 today is 25% of the $1.50/$7.50 rate that takes effect in January. If your workload doesn’t need a real-time response, routing it through the batch endpoint now captures a discount that gets relatively larger, not smaller, as the standard rate rises. See our batch API pricing breakdown for how this compares across providers.
Benchmarks and capability improvements
Google’s own materials describe gains in coding, document analysis, and workflow automation over 3.6 Flash. We have not independently verified specific benchmark scores for 3.7 Flash against 3.6 Flash — at the time of writing, Google’s release materials describe the improvement areas in general terms rather than publishing a side-by-side benchmark table the way it did for the 3.5 Flash launch. Treat any specific percentage figures you see elsewhere for 3.7 Flash as unverified until Google publishes a comparable benchmark table for this specific model.
Coding and agent relevance
Flash-tier models in the Gemini 3.x line have consistently targeted agentic and tool-use workloads rather than competing purely on reasoning benchmarks against Pro-tier models. Google’s description of 3.7 Flash as built for “agentic workflows and multimodal reasoning” continues that positioning. If you’re routing model calls by task complexity — a cheap Flash-tier model for high-volume, well-defined steps and a Pro-tier model for harder reasoning — 3.7 Flash is a direct drop-in candidate for the Flash side of that split, at least through the end of 2026 pricing window.
Gemini 3.7 Flash vs Gemini 3.6 Flash
| Gemini 3.6 Flash | Gemini 3.7 Flash | |
|---|---|---|
| Release date | July 21, 2026 | August 13, 2026 |
| Standard input/output (per 1M) | $1.50 / $7.50 | $0.75 / $3.75 (through Dec 31, 2026) |
| Pricing structure | Standard, ongoing | Introductory, confirmed to double Jan 1, 2027 |
| Positioning | Computer use, token efficiency | Agentic workflows, multimodal reasoning |
The headline difference right now is price: 3.7 Flash costs half of 3.6 Flash’s standard rate, but only until the end of 2026. After January 1, 2027, 3.7 Flash’s standard rate ($1.50/$7.50) lands at exactly what 3.6 Flash already charges today. Whether 3.7 Flash is a straightforward upgrade over 3.6 Flash depends on capability gains Google hasn’t yet published in detail — see the benchmarks caveat above.
Who should use it
Teams running high-volume agentic or multimodal workloads who can take advantage of the discounted rate before it expires, and who are comfortable re-evaluating cost once the January 2027 price change lands. It’s a reasonable default for new projects launching now, provided you build the pricing step-up into your cost model rather than treating today’s rate as permanent.
Limitations
- No independently verified benchmark comparison against 3.6 Flash or competing models at time of writing.
- The attractive headline price is confirmed to expire January 1, 2027 — budget accordingly.
- As with other recent Gemini releases, treat vendor-reported capability claims as vendor-reported until third-party benchmark results are available.
For the full lineup of Flash-tier and other Gemini models, see our Gemini 3.6 Flash guide and AI model comparisons.