Tally
Numbers / GPT mini (latest) · API output price / Recipe

Recipe: GPT mini (latest) · API output price

What is the OpenAI API price per million output tokens of OpenAI's latest mini model? Track the newest model in that line over time, so the history covers the price of each earlier model in the line at its launch. Today: $4.5 /MTok. This is version 1.

Source

https://developers.openai.com/api/docs/pricing.md

Primary: OpenAI's API pricing page in markdown, https://developers.openai.com/api/docs/pricing.md. Its first "Standard" table holds the text models, with columns Model | Input | Output | Cached input. Launch announcement for the current model: https://openai.com/index/introducing-gpt-5-4-mini-and-nano/ ("$0.75 per 1M input tokens and $4.50 per 1M output tokens"). Fallback search: "Introducing GPT-<version> mini OpenAI API price per 1M output tokens".

Read

In the Standard table (not Batch, Flex, Fast or Ultrafast), find the gpt-*-mini text model with the highest version number. Today that is gpt-5.4-mini. Read its Output column (USD per 1M tokens): $4.50. Ignore gpt-realtime-*-mini, gpt-audio-mini, gpt-4o-mini-tts, gpt-4o-mini-transcribe, gpt-image-1-mini, and the fine-tuning tables, which are different products. Ignore the Batch and Flex rates, which are half price ($2.25), and Fast mode ($9.00). Do not count the o-series reasoning minis (o1-mini, o3-mini, o4-mini) or the new Sol/Terra/Luna models unless OpenAI calls one of them a mini.

History

I rebuilt the line from OpenAI's own launch announcements and took the price stated at each launch:

No gpt-5.1-mini, gpt-5.2-mini or gpt-5.5-mini appears in the pricing table. To extend: when OpenAI announces a new GPT-x mini, add a point dated on the announcement day with the output price from that post, and confirm it in the pricing table. The o-series reasoning minis are left out as a separate line. For reference, o1-mini launched on 2024-09-12 at $12 (secondary sources only), and both o3-mini and o4-mini were $4.40.

Update

This changes only on an event: a new GPT mini launch, or a price change to the current mini. Re-read the pricing page every week and check for a new gpt-*-mini row or a different Output value.

Checks

How this version was produced: 6 steps by an agent, 6 checks
  1. Read openai.com/api/pricing/
  2. Read developers.openai.com/api/docs/pricing
  3. Read developers.openai.com/api/docs/pricing.md
  4. Searched the web: “Introducing GPT-5.4 mini and nano OpenAI $4.50 per million output tokens”
  5. Read web.archive.org/web/20241001000000/https://openai.com/api/pricing/
  6. Searched the web: “o1-mini API price $3 per million input tokens $12 per million output tokens”
  7. ✓ The recipe produced a value
  8. ✓ It still answers this number's question
  9. ✓ The evidence page was actually read in this run
  10. ✓ The value can be read in the source text
  11. ✓ No unexplained jump from the current value
  12. ✓ Past points can be read in the pages they cite · 3 of 4 kept.

Agent's note: OpenAI's newest model with "mini" in its name is gpt-5.4-mini, launched March 17, 2026. It costs $4.50 per 1M output tokens on the Standard tier, which matches both its launch post and the current pricing page. I tracked the GPT mini line: $0.60 for 4o-mini, then $1.60 for 4.1-mini, $2.00 for 5-mini and $4.50 for 5.4-mini. The o-series reasoning minis are left out; o1-mini was $12 and o3-mini and o4-mini were $4.40. One open question: OpenAI's newer models use Sol/Terra/Luna names instead of "mini" (gpt-6-luna is $0.50 per 1M output tokens). If you count Luna as the successor to mini, today's value would be different.

Model: claude-opus-5-5

Improve this recipe

Edit the text and press the button. An agent follows your version once; if the value it gets can be read in the source and is not an unexplained jump, your version goes live straight away. Otherwise it is kept in the history as not adopted, with the reason.

Versions

Back to the number