GPT-5.4 mini in OpenAI Codex
OpenAI - Lightweight
In OpenAI Codex, GPT-5.4 mini meters at 18.75 credits per 1M input tokens and 113 per 1M output tokens. At $0.04 per credit that is exactly $0.75 and $4.50 per 1M tokens, the same as calling the OpenAI API directly. A typical daily-coding request works out to about 0.3 credits (about $0.01). Verified against OpenAI's Codex rate card on 2026-07-21.
Credit rate card
- Credits / 1M in
- 18.75
- Credits / 1M cached
- 1.88
- Credits / 1M out
- 113
- Cost per credit
- $0.04
Credits per request
Defined workload baskets, not measured tasks.
| Workload | Credits | Cost | Direct API cost |
|---|---|---|---|
| Short chat turn | 0.2 | $0.01 | $0.01 |
| Daily coding request | 0.3 | $0.01 | $0.01 |
| Agent session step | 1.0 | $0.04 | $0.04 |
At about 0.3 credits per daily-coding request, GPT-5.4 mini ranks #1 of 4 current-generation models by cost, about 1x the cheapest (GPT-5.4 mini).
Credit-to-API parity
The implied dollar value of a credit spent on GPT-5.4 mini, backed out from the direct API rate card divided by the credit rate card. It should land on $0.04 for both input and output.
- Implied $ / credit, input
- $0.04
- Implied $ / credit, output
- $0.04
- Markup at the daily workload
- 0.26%
- Ratio match
- Rounded
Input : cached : output ratio in credits is 150 : 15 : 904, versus 10 : 1 : 60 in direct API dollars.
The credit card rounds output to a whole 113 credits per 1M tokens (the exact API-parity value is 112.5). At $0.04 per credit that reads as $4.52 versus the $4.50 API rate, a 0.4% rounding artifact.