Guides / gemini 2.5 flash pricing
Gemini 2.5 Flash pricing and the eight-to-one output spread
Gemini 2.5 Flash pricing is $0.30 per million input tokens and $2.50 per million output tokens, a spread of roughly eight to one. That makes answer length, not prompt length, the lever that moves your bill. Culpa, a local-first LLM cost, margin, and forecast ledger, splits input and output per call so you can see which side is growing.
Why this happens
Flash carries the widest practical output-to-input spread of the workhorse tier. Input is close to free at $0.30 per million, so stuffing context is cheap, while every output token costs more than eight input tokens. The optimisation that works everywhere else, trimming the prompt, barely moves this bill. The one that works here is constraining the response format, and most teams try them in the wrong order.
What this usually looks like
- Trimming your prompt produced almost no change in spend.
- Output tokens are a small fraction of your token count and most of your bill.
- Responses are open-ended prose where a structured, short answer would do.
- Cost per call varies widely between requests, tracking response length rather than input size.
Free, no card, no account
Run the free Cost Leak Scan
It shows your most expensive conversation before you install anything.
Mistakes that cost the most
| Mistake | Why it hurts | Do instead |
|---|---|---|
| Optimising the prompt first. | Input is $0.30 per million here, so halving a long prompt saves very little. | Constrain the output first. Set max tokens and request a compact format, then measure. |
| Asking the model to explain its reasoning in the response. | Narration is output tokens at eight times the input rate, and it's usually not read. | Ask for the answer only. If you need reasoning, request it on a sampled subset rather than every call. |
| Assuming a cheap input rate means a cheap model. | A verbose Flash response can cost more than a terse response on a nominally pricier model. | Compare models on your own output-to-input ratio, not on the input rate you noticed first. |
Run this check tonight
- Split your Flash spend into input and output. The output share is usually the surprise.
- Measure median and 90th percentile output tokens per call. The gap tells you how uncontrolled the format is.
- Check whether max tokens is set at all. On an eight-to-one spread it's the cheapest control you have.
- Take your ten longest responses and ask whether any of that length was read by anything downstream.
- Recompute your bill with output halved. That's the size of the prize before you touch the prompt.
Halving the prompt against halving the answer
Illustrative example
A summarisation feature sending 50 million input and 6 million output tokens a month on Gemini 2.5 Flash.
Both edits save the same $7.50, but one means removing 25 million tokens and the other means removing 3 million. On this model the output side is 8.3 times more efficient to work on.
Every number, with its confidence and source
| Figure | What it means | Confidence | Source |
|---|---|---|---|
| $0.30 / $2.50 per million | Gemini 2.5 Flash input and output rates | calculated | Culpa price book row for gemini/gemini-2.5-flash, $0.0003 and $0.0025 per 1k, effective 2026-07-02, verified against Gemini API pricing. |
| 8.3x | output rate as a multiple of the input rate on Gemini 2.5 Flash | calculated | $2.50 divided by $0.30, from the same price-book row. |
What a generic answer can’t know
A rate card gives you the two numbers. It can't tell you your own output-to-input ratio, which decides which lever is worth pulling, or which feature produces the long responses. That needs per-call token splits joined to your features. Culpa records them on your infrastructure, keeps your prompts and responses there, and counts the calls to run your plan.
Questions founders ask next
How much does Gemini 2.5 Flash cost?
$0.30 per million input tokens and $2.50 per million output, verified 2026-07-02 against the Gemini API pricing page. Cached input runs at $0.03 per million, a tenfold discount on the input side.
Why does trimming my prompt barely change my Flash bill?
Because input is only $0.30 per million. On an eight-to-one spread the money is on the output side, so response length is where the saving is. Prompt trimming is the right instinct on most models and the wrong one here.
Is Flash cheaper than Gemini 2.5 Pro?
Substantially. Pro bills $1.25 input and $10.00 output per million against $0.30 and $2.50 here, so Flash is roughly four times cheaper on both sides. Whether that trade works depends on your task, so test on your own data.
On your infrastructure
Culpa runs on your infrastructure. Your prompts and responses never leave it. Culpa counts calls to run your plan, and it fails open, so if it ever breaks your app keeps running.
Free, no card, no account
Run the free Cost Leak Scan
It shows your most expensive conversation before you install anything.
Keep reading
Sources: Gemini API pricing. Last reviewed 2026-08-01, rates effective 2026-07-02. Plain text version.