GPT-4o mini API pricing at a glance
As of November 9, 2026, OpenAI’s pricing page lists GPT-4o mini’s standard API rates in USD per 1 million tokens: $0.15 for input, $0.075 for cached input, and $0.60 for output. These are standard-processing prices, not Batch prices or a ChatGPT subscription fee. (developers.openai.com)
| Token type | Standard API price (USD per 1M tokens) |
|---|---|
| Input | $0.15 |
| Cached input | $0.075 |
| Output | $0.60 |
The pricing page separates its Standard pricing data from its Batch pricing data. The figures above and the example below use standard pricing; they are not a Batch estimate. Batch requests have a separate pricing schedule, so use that schedule if you are budgeting for Batch processing. (developers.openai.com)
OpenAI’s model page lists gpt-4o-mini and snapshot gpt-4o-mini-2024-07-18, describes text and image input with text output, and gives a 128,000-token context window and 16,384-token maximum output. The same page marks the model deprecated. Check current account availability and migration guidance before using it for a new production integration. (developers.openai.com)
How to estimate your standard API cost
Estimate input and output separately. Count cached input separately only when it is reported as cached. Using OpenAI’s listed standard API rates:
Estimated standard token cost = (uncached input tokens ÷ 1,000,000 × $0.15) + (cached input tokens ÷ 1,000,000 × $0.075) + (output tokens ÷ 1,000,000 × $0.60).
Example: Suppose one month of standard API usage has 10 million uncached input tokens and 2 million output tokens, with no cached input included:
- Input: 10 × $0.15 = $1.50
- Output: 2 × $0.60 = $1.20
- Estimated standard token cost: $2.70
This is arithmetic using the standard rates above—not a forecast, a typical monthly bill, or a guaranteed charge. It excludes cached input and separately priced features. It does not apply to Batch processing; use the separate Batch pricing schedule for that tier. (developers.openai.com)
For a per-1,000-token conversion, the standard rates are $0.00015 per 1,000 input tokens and $0.0006 per 1,000 output tokens. For a monthly estimate, base token volumes on representative traffic, including conversation history if your application resends it.
Cached input: a conditional lower rate
OpenAI lists cached input at $0.075 per million tokens under standard pricing. Its prompt-caching explanation says caching applies to eligible repeated prompt prefixes; it describes cached tokens beginning at 1,024 tokens, with additional cached tokens counted in 128-token increments. Don’t assume every request—or every input token—will receive the cached rate. (developers.openai.com)
Production budgeting: check status before committing
OpenAI’s API model page marks GPT-4o mini as deprecated while still displaying its prices and model details. If you already use it, check your account and OpenAI’s current model documentation for availability and migration guidance. For a new production workload, compare a currently supported model and test representative prompts if quality or latency could change your choice. (developers.openai.com)
Before setting a budget:
- Confirm the model’s current status and availability for your API account.
- Measure input and output tokens from representative requests, including conversation history your application resends.
- Apply the cached-input rate only to tokens actually reported as cached.
- Use standard or Batch rates consistently with the processing tier you plan to use, and recheck the applicable schedule before forecasting spend.