Input
- Official OpenAI
- $2.5
- Here, from (−60%)
- $1
- Here, best (−70%)
- $0.75
GPT-5.4 is the proven balanced tier of the previous generation — a workhorse for coding and production pipelines, at the same list price as GPT-5.6 Terra.
| Rate | Official OpenAI | Here, from (−60%) | Here, best (−70%) |
|---|---|---|---|
| Input | $2.5 | $1 | $0.75 |
| Cached input | $0.25 | $0.1 | $0.075 |
| Cache write | $2.5 | $1 | $0.75 |
| Output | $15 | $6 | $4.5 |
Every request is metered at the official rate first, then your progressive B2C discount (60% at the start, up to 70% as cumulative top-ups grow) is subtracted before it touches your prepaid balance. Context window: 272K tokens. Max output: 32K tokens. Reasoning efforts: none, low, medium, high, xhigh.
Create a free account, generate one key, and point any OpenAI-compatible tool at https://openai.api.apitoken.sale/v1 with model ID gpt-5.4 — Responses and Chat Completions both work, authenticated with Authorization: Bearer. New accounts include $10 of API usage at official prices — enough to test the model before topping up.
Officially $2.50 per 1M input tokens and $15 per 1M output tokens. With the apiToken.sale discount that starts at $1/$6 and reaches $0.75/$4.50.
They share a list price, and Terra is the newer balanced tier — prefer it for new projects. Stay on 5.4 when your prompts and evals are pinned to it.
gpt-5.4. One apiToken.sale key and balance covers it alongside every other Claude and GPT model.
Run GPT-5.4 on the OpenAI-compatible API at up to 70% off — instant key, prepaid balance, card or crypto.