Skip to main content

OpenAI pricing guide

OpenAI API Pricing Explained: Tokens, Cache, Batch, and Cost

OpenAI products use several billing units. Text generation separates input and output tokens, embeddings price input volume, image generation varies by output settings, and transcription can be planned by audio duration.

Formula-driven examplesSource-linked pricing snapshots

Separate standard, cached, and batch usage

Standard input is the normal prompt rate. Eligible cached input can be cheaper, while batch processing can offer different rates in exchange for asynchronous completion. Do not apply a discount unless the workload actually qualifies.

For chat and agent workloads, output tokens can dominate because their published rate is often higher. Track the input-to-output ratio in production rather than optimizing only prompt size.

monthly text cost = request cost × requests per day × 30

Use the billing unit for each modality

An image estimate should preserve model, size, and quality. An embedding estimate should include re-embedding frequency, and transcription should include total audio minutes plus any provider rounding rule.

Worked example

GPT-4.1 mini: 1M input tokens + 1M output tokens

Using the versioned rates below, this example workload is estimated at $2.00. This isolates provider usage only and does not include taxes, regional premiums, retries, storage, network traffic, or unrelated infrastructure.

Current pricing references

These versioned records support the examples above. Check the date and provider source before using them in a production forecast.

Provider / modelInput or unitOutputStatusSource

OpenAI

GPT-4.1 mini

$0.40 per 1M tokens$1.60 / 1MVerifiedOpenAI model pricing

Checked Jun 21, 2026

OpenAI

GPT-4.1

$2.00 per 1M tokens$8.00 / 1MVerifiedOpenAI model pricing

Checked Jun 21, 2026

OpenAI

text-embedding-3-small

$0.02 per 1M tokensVerifiedOpenAI embedding model pricing

Checked Jun 21, 2026

Frequently asked questions

Is the OpenAI API a fixed monthly subscription?

API billing is generally usage based and separate from consumer ChatGPT subscriptions. Verify account-specific terms and service tiers.

Should I include cached input in every estimate?

No. Use standard input pricing until you have measured cache eligibility and hit rate for the production prompt pattern.

Related calculators and guides