A8gent
GlossaryAgent concepts

What is gemini 2.5 Flash-Lite?

Also called: gemini-2.5-flash-lite, gemini flash lite pricing, cheapest gemini model

Gemini 2.5 Flash-Lite is the cheapest model ID on Google's Gemini API pricing page at $0.10 per million input tokens and $0.40 per million output tokens, which makes it the default for bulk classification and extraction work.

Written by A8gent
Contents
  1. At a glance
  2. What Gemini 2.5 Flash-Lite is
  3. What ten cents a million buys
  4. Where it is the wrong tool
  5. What you should know
  6. FAQ
  7. Related terms

At a glance

Model ID
gemini-2.5-flash-lite
Input price
$0.10 per 1M tokens for text, image and video (checked August 2026)
Output price
$0.40 per 1M tokens
Free tier
Listed as free of charge, subject to free-tier limits
Versus gemini-2.5-flash
A third of the input price and roughly a sixth of the output price
Position
Cheapest published row on Google's Gemini API pricing page
Best suited to
Bulk classification, extraction, routing and structured output at volume

What Gemini 2.5 Flash-Lite is

gemini-2.5-flash-lite is Google's cheapest published model ID. As listed on the Gemini API pricing page in August 2026, the paid-tier rate is $0.10 per million input tokens for text, image and video, and $0.40 per million output tokens. It is also listed as free of charge on the free tier, subject to that tier's rate limits.

Flash-Lite is Google's naming for the tier below Flash, and the pricing gap is significant: gemini-2.5-flash is $0.30 in and $2.50 out, so Flash-Lite is a third of the input price and roughly a sixth of the output price of the tier above it.

What ten cents a million buys

The number is worth making concrete because it changes what is worth automating. At $0.10 per million input tokens, processing an archive of 200,000 documents averaging 1,000 tokens each is 200 million tokens, or $20.00 of input. Adding a 50-token structured label to each is 10 million output tokens, or $4.00.

Twenty-four dollars to classify 200,000 documents is the kind of figure that makes "classify everything" a reasonable default rather than a project. Bulk triage, deduplication, language routing, first-pass tagging and structured extraction into JSON all fall into this band.

Where it is the wrong tool

Anywhere the output space is open and errors compound. Multi-step reasoning, choosing between many similar tools, and anything a customer reads unreviewed are jobs for a larger model or for a small model behind a guardrail. The right pattern is usually a cheap first pass that handles the easy majority and escalates the uncertain remainder - which requires the model to be able to say it is uncertain, so build that into the schema.

What you should know

Compare it against gpt-5.6-luna at $0.20 in and $1.20 out on your own task rather than on price alone. Flash-Lite is half the input price and a third of the output price, but a model that gets 4% more of your labels right can be worth several times the difference once you count the cost of a wrong label reaching a customer.

Note also that the 2.5-series Flash models are cheaper than the newer gemini-3.7-flash even at its promotional rate. On Google's lineup, moving to a newer generation is a capability decision, not a cost saving.

Prices checked August 2026 against Google's pricing page. Verify before budgeting.

Gemini 2.5 Flash-Lite FAQ

How much does gemini-2.5-flash-lite cost?

$0.10 per million input tokens for text, image and video, and $0.40 per million output tokens, as listed on Google's Gemini API pricing page in August 2026. It is also available free of charge on the free tier, subject to that tier's rate limits.

What is the cheapest Gemini model?

Gemini 2.5 Flash-Lite is the cheapest row on Google's published pricing page at $0.10 input and $0.40 output per million tokens. Gemini 2.5 Flash sits above it at $0.30 and $2.50, and the newer 3.7 Flash is more expensive than both.

What is Gemini 2.5 Flash-Lite good for?

High-volume, narrow-output work: classification into fixed labels, extraction into a schema, routing, deduplication, language detection and first-pass tagging. At ten cents a million input tokens, classifying an entire archive costs tens of dollars rather than thousands.

Is Flash-Lite cheaper than gpt-5.6-luna?

On list price yes - $0.10 against $0.20 input and $0.40 against $1.20 output. Whether it is cheaper in practice depends on accuracy on your task, because a model that mislabels more often costs you in downstream errors. Test both on the same eval set.

When should I not use Flash-Lite?

When the output space is open and mistakes compound: multi-step reasoning, tool selection among many similar tools, and anything a customer reads without review. Use it as a cheap first pass that escalates uncertain cases, and build an uncertainty signal into the output schema so it can.

Primary source

Google - Gemini API pricing

Checked August 2026. Anything priced or rate-limited moves, so verify against the source before you budget.

Want this built rather than explained?

Knowing what gemini 2.5 flash-lite means is one thing; having it running in your business is another. We scope one workflow at a time, quote a fixed price in writing after a free consult, build it, and hand it over with documentation you own.