OpenAI logoOpenAIopenai.com

Proprietary

GPT-4o mini

GPT-4o mini is OpenAI's low-cost workhorse for simple, high-volume text and image jobs. If you run the same small task thousands or millions of times, this is the model to reach for.

Context window
128,000 tokens
Max output
16,384 tokens
Released
18 Jul 2024
Knowledge cutoff
Oct 2023
API identifier
gpt-4o-mini-2024-07-18
Modalities
Vision · Text · Image · Video · Audio (planned for future)

Pricing · per 1M tokens

Input$0.15
Output$0.60

Overview

What GPT-4o mini is good at

Think tagging products, sorting support tickets, or pulling fields out of documents. It's fast, it costs very little per call, and it reads both text and images. At scale, that low price per call is the whole point.

The context window holds 128,000 tokens, so long documents fit fine. Each response can run up to 16,384 tokens.

Now the honest part. GPT-4o mini is a legacy model, and OpenAI's frontier line has moved on to GPT-5.x, including GPT-5.4-mini. For work that needs real reasoning or careful judgment, step up to a newer model.

But for simple, repeat jobs at huge volume, it's hard to beat on cost. Point it at the boring, high-count work. Save the smarter models for the hard problems.

It launched in July 2024, with knowledge current to October 2023.

Real numbers

What GPT-4o mini costs in practice

Real enterprise workloads, with the token assumptions shown openly. Your mileage varies — these are honest starting points, not guesses.

E-commerceSales & Marketing

Tag a huge product catalog

$0.00014per request
$105per month

You've got hundreds of thousands of products, and every one needs clean category tags and attributes. GPT-4o mini reads each title and description, then returns structured tags you can drop straight into your catalog. It won't win awards, but it does the boring work fast and for very little.

450 input tokens120 output tokens750,000 requests / month

Outputs stay short, so the per-product cost stays tiny and the whole catalog run is a small line item.

FinanceFinance & Accounting

Pull fields from invoices

$0.00027per request
$40.5per month

Your accounts team drowns in invoices every month. GPT-4o mini reads each one and pulls the vendor, date, line items, and total into clean fields. Let it handle the simple 90%, and send the messy ones to a person.

1,000 input tokens200 output tokens150,000 requests / month

Invoices make the input a bit longer, but the per-document cost stays low — volume is where it pays off.

RetailCustomer Service

Sort and route support tickets

$0.000063per request
$25.2per month

Your support inbox fills up faster than the team can sort it. GPT-4o mini reads each ticket and tags the intent, the urgency, and the right queue. It's first-line triage, so a human only touches what actually needs them.

300 input tokens30 output tokens400,000 requests / month

Labels are only a few tokens each, so triage is about as low-cost per ticket as it gets.

How we work these out: cost = (input ÷ 1M × $0.15) + (output ÷ 1M × $0.60), then × monthly volume. List prices only, no cached-input discount applied — so these are the ceiling, not the floor.

FAQ

Common questions about GPT-4o mini

How much does GPT-4o mini cost?+

GPT-4o mini costs $0.15 per million input tokens and $0.60 per million output tokens. That's OpenAI's list price. Output is priced four times higher than input, so shorter responses keep your bill down. For simple, high-volume jobs, it's one of the lower-cost options around.

GPT-4o mini vs GPT-4o — what's the difference?+

GPT-4o mini is the smaller, lower-cost sibling built for speed and volume on simple tasks. GPT-4o is the larger model, better at hard reasoning and nuance, but it costs more per call. Reach for mini on high-count simple work, and use GPT-4o when the task needs more thinking.

What is GPT-4o mini's context window?+

GPT-4o mini has a 128,000-token context window. That's roughly a long book's worth of text in a single request. It can also write back up to 16,384 tokens per response, so you can feed it big documents and still get a full answer.

Is GPT-4o mini good for classifying support tickets at high volume?+

Yes — ticket classification is exactly what GPT-4o mini is built for. It's fast, the cost per ticket is tiny, and sorting or tagging doesn't need deep reasoning. Run it as first-line triage, then pass the tricky cases to a human or a stronger model.

GPT-4o mini vs GPT-5.4-mini — which should you use?+

GPT-5.4-mini is the newer model, and GPT-4o mini is the older, lower-cost one. If you need better reasoning or more current knowledge, GPT-5.4-mini wins. If you're running simple, high-volume jobs and want the lowest cost, GPT-4o mini still holds up. Many teams keep it for the bulk work and use newer models where thinking matters.

Pricing as of 8 Jul 2026 · reviewed monthlySources: OpenAI API pricing · OpenAI model docs

Not sure GPT-4o mini fits your workload?

We'll size the right model for your use case and give you a real cost estimate. One free call, no slides.

Talk to us