Gemini 3.7 Flash: Pricing, Context and Uses

Superseded by Gemini 3.8 FlashGoogle's current speed-tier model in Cursor at $0.75 per million input tokens and $3.75 per million output; it succeeds Gemini 3.7 Flash. Press Enter for the full definition.. Cursor's model page now calls Gemini 3.7 FlashGoogle's previous speed-tier model in Cursor, hidden by default and succeeded by Gemini 3.8 Flash at the same published rates. Press Enter for the full definition. Google's previous speed-tier model and recommends Gemini 3.8 Flash, Google's latest, at the same price. Cursor's pricing index marks 3.7 Flash hidden by default. Start new work from the current model guide, and keep this page for predecessor comparisons.

The model ID is gemini-3.7-flash. Cursor says the model pairs strong agent-based coding performance with low per-token costs, and rates it fast for speed, low for cost and high for intelligence. As a third-party model, it bills from prepaid credit on The X SubscriptionThe subscription Cursor is sold through: one plan for Grok, Cursor and X, with Free, Plus, Premium, Super and Ultra tiers, per-seat Teams, a weekly included pool for Cursor models shared with Grok, and third-party models billed at API list price from prepaid credit. Press Enter for the full definition. or from the Other ModelsOn legacy Cursor plans, the monthly pool for supported third-party models from providers such as Anthropic, Google and OpenAI, separate from Cursor Models. On The X Subscription third-party models are billed from prepaid credit instead. Press Enter for the full definition. pool on a legacy plan, not from the included pool that covers Grok and ComposerCursor's own fast coding model, tuned for the editor and priced well below frontier models; the recommended day-to-day model for executing a plan. Press Enter for the full definition..

What the release offered was price plus context: fresh input at $0.75 per million tokens, a cached read at $0.075, and a card that lists 200k as the standard context with a 1M maximum. A long session can keep a large working set in view without paying a frontier-model output rate on every response.

For new work, use Gemini 3.8 FlashGoogle's current speed-tier model in Cursor at $0.75 per million input tokens and $3.75 per million output; it succeeds Gemini 3.7 Flash. Press Enter for the full definition., which costs the same and is the Flash model Cursor recommends. Gemini 3.7 FlashGoogle's previous speed-tier model in Cursor, hidden by default and succeeded by Gemini 3.8 Flash at the same published rates. Press Enter for the full definition. is still the right pick when its identity matters: you need to reproduce a result it produced, measure a move to 3.8, or work under an approval that names gemini-3.7-flash.

When you do run it, keep the task bounded. A repository search followed by several related changes, a short investigation of a failing test, or a frontend fix checked in the browser all suit a fast model with reasoning. Ask it to inspect before editing, name the command or screen state that proves the work is done, and require the relevant tests afterward.

The low output price helps when a task takes several tool calls, but a cheap failed attempt still creates review work. Our default is a reviewable task that can finish inside one focused session.

Cursor lists Gemini 3.7 FlashGoogle's previous speed-tier model in Cursor, hidden by default and succeeded by Gemini 3.8 Flash at the same published rates. Press Enter for the full definition. at $0.75 per million input tokens, $0.075 per million cache reads and $3.75 per million output. Gemini 3.8 FlashGoogle's current speed-tier model in Cursor at $0.75 per million input tokens and $3.75 per million output; it succeeds Gemini 3.7 Flash. Press Enter for the full definition. lists the same three rates. The cache-write column is blank, so that rate is unpublished; don't read the blank as a claim that cache writes are free.

Access splits by plan family. The X SubscriptionThe subscription Cursor is sold through: one plan for Grok, Cursor and X, with Free, Plus, Premium, Super and Ultra tiers, per-seat Teams, a weekly included pool for Cursor models shared with Grok, and third-party models billed at API list price from prepaid credit. Press Enter for the full definition. offers third-party models from Premium upward, paid from prepaid credit at these rates, and leaves them out of Plus and Free. Among legacy plans, Pro, Pro+, Ultra and Teams include the Other ModelsOn legacy Cursor plans, the monthly pool for supported third-party models from providers such as Anthropic, Google and OpenAI, separate from Cursor Models. On The X Subscription third-party models are billed from prepaid credit instead. Press Enter for the full definition. pool the model draws from, while Cursor StartAn India-only legacy individual plan at ₹649 a month centered on Cursor's own models, with cloud and mobile access but without Other Models, Auto, Bugbot, Automations or the Agent SDK; Cursor names Plus as the comparable plan on The X Subscription. Press Enter for the full definition. and Hobby don't. See Cursor pricing for how each family bills.

Teams and Enterprise add the Cursor Token Rate of $0.25 per million tokens to eligible third-party requests, on input, output and cached tokens alike. Gemini 3.7 FlashGoogle's previous speed-tier model in Cursor, hidden by default and succeeded by Gemini 3.8 Flash at the same published rates. Press Enter for the full definition. is on Cursor's excluded list, so a team pays the model rate alone.

Check the usage record and the contract before you turn a rate row into a team budget.

Treat 200k as the normal working window and 1M as the documented ceiling. A million-token maximum can hold a large repository slice, but it doesn't make every file relevant. Begin at the feature boundary, search outward when imports or test failures point elsewhere, and keep the acceptance check in view.

The 90% cache-read discount pays off when the same context comes back. A first pass over one million fresh input tokens costs $0.75 at the published rate, and a later cached read of the same amount costs $0.075. That saving suits a sustained codebase session far better than a run of unrelated prompts.

Because the cache-write rate is blank, check the actual usage record instead of estimating a whole session from cache reads.

1. 1Open a fresh session for one task and name the expected output and verification command.
2. 2Load the smallest repository slice that explains the behavior, then search for missing dependencies.
3. 3Keep stable instructions and relevant files in the same session so repeated context can be cached.
4. 4Check the context ring next to the prompt input before adding broad logs, generated files, or unrelated folders.
5. 5Inspect the usage breakdown after the run before using its cost as a team estimate.

A larger window helps most when the work is connected. If the next request has a different goal, start another session, so each review stays smaller and you stop carrying context that no longer helps. The Cursor model guide applies one rule across models: choose a default by accepted changes and total review cost.

Against Gemini 3.8 FlashGoogle's current speed-tier model in Cursor at $0.75 per million input tokens and $3.75 per million output; it succeeds Gemini 3.7 Flash. Press Enter for the full definition., the published numbers match: the same three rates, the same 200k standard context with a 1M maximum, and the same fast, low and high ratings. The 3.8 page adds three effort levels that the 3.7 page doesn't list, and Cursor publishes no quality comparison between them. Against Gemini 3.6 Flash, 3.7 halved every published rate: input fell from $1.50 to $0.75, cache reads from $0.15 to $0.075, and output from $7.50 to $3.75.

Cursor places Gemini 3.1 ProGoogle's frontier model in Cursor for work that combines images and code, especially UI implementation and visual code analysis. Press Enter for the full definition. on a different trade-off: medium for speed and cost, frontier for intelligence, at $2 input, $0.20 cache read and $12 output per million tokens.

When the choice matters, run the same small task in fresh sessions, with the prompt, files and test held fixed. Compare the final diff, failed attempts, elapsed time and usage, because the rate card can't tell you which result holds up on your repository.

Gemini 3.7 FlashGoogle's previous speed-tier model in Cursor, hidden by default and succeeded by Gemini 3.8 Flash at the same published rates. Press Enter for the full definition. gets the full agent tool set. Cursor's model page lists file and folder search, web search, reading and editing files, shell commands, browser control, image generation, clarifying questions and rule fetching. The product surface, sandbox, approvals, credentials and team settings still decide what a particular run can do.

Write the verification step into the prompt: run the named test, inspect the target route, or show the final file. With browser and shell access the model can collect better evidence than a chat reply, but only when the task asks for it. External writes and production changes still need the authority that applies to that system.

Not with the pool Cursor lists today. Cursor Router, available on Teams and Enterprise, picks a model for each Balance or Intelligence request in Auto from a pool that changes over time, and you can't hand-pick the model it uses. Cursor's help page lists the models in that pool, and none of them is a Gemini model.

Billing follows the mode, and Cursor's own pages disagree on Cost. Its docs say Cost mode bills as a Cursor-model request, while its Cursor Router help page says all Auto modes bill at the routed model's list price. Balance and Intelligence charge the list price of the model each request is routed to, and a third-party route adds the Cursor Token Rate unless the model is on the excluded list.

Select Gemini 3.7 FlashGoogle's previous speed-tier model in Cursor, hidden by default and succeeded by Gemini 3.8 Flash at the same published rates. Press Enter for the full definition. directly when the model identity is part of the test or the approval.