AI Rate CardGPT-5.5 and GPT-5.4 API pricing: Pro, mini and nano at every provider
Updated Nov 4, 2026
GPT-6 is OpenAI's newest line, but GPT-5.5 and the GPT-5.4 family are still what many applications run on. OpenAI sells them itself, and Azure, routers and resellers sell them too - up to 30 providers per model. The tables below are rebuilt from today's prices on every page view.
The models in short
OpenAI's own price per 1 million tokens, input / output, in the 23 October 2026 collection:
- GPT-5.5 - $5 / $30, cached input $0.50; about 1 million tokens of context.
- GPT-5.4 - $2.50 / $15, cached input $0.25.
- GPT-5.4 Pro - $30 / $180: the slow, most thorough version; no cached-input discount.
- GPT-5.4 mini - $0.75 / $4.50, 400 thousand tokens of context.
- GPT-5.4 nano - $0.20 / $1.25, the cheapest of the family. Most other providers keep OpenAI's price; a few list less - check the card, as a price far below OpenAI's is often a promotion, a batch price or marked for review. For the newest models see GPT-6 pricing. For the new GPT-6.1 Terra ($1 / $5) see GPT-6 pricing. (GPT-6.1 Terra)
GPT-5.5: every provider
| Provider | Model | Input, $ | Output, $ |
|---|---|---|---|
| xpersona | GPT-5.5 | $0.001 / request | $12.00 |
| openrouter | OpenAI: GPT-5.5 (batch) | $2.50 | $15.00 |
| requesty | GPT-5.5 | $2.50 | $15.00 |
| vercel | GPT 5.5 (Batch) | $2.50 | $15.00 |
| cometapi | GPT-5.5 | $4.00 | $24.00 |
| meganova | GPT-5.5 | $4.00 | $24.00 |
| poe | GPT-5.5 | $4.55 | $27.27 |
| jiekou | gpt-5.5 | $4.75 | $28.50 |
| openai | GPT-5.5 | $5.00 | $30.00 |
| 302ai | GPT-5.5 | $5.00 | $30.00 |
| abacus | GPT-5.5 | $5.00 | $30.00 |
| aihubmix | GPT 5.5 | $5.00 | $30.00 |
| azure | GPT-5.5 | $5.00 | $30.00 |
| daoxe | GPT-5.5 | $5.00 | $30.00 |
| digitalocean | GPT-5.5 | $5.00 | $30.00 |
Live data: this table is built from today's prices.
GPT-5.4 mini: every provider
| Provider | Model | Input, $ | Output, $ |
|---|---|---|---|
| openrouter | OpenAI: GPT-5.4 Mini (batch) | $0.38 | $2.25 |
| requesty | GPT-5.4 mini | $0.38 | $2.25 |
| vercel | GPT 5.4 Mini (Batch) | $0.38 | $2.25 |
| xpersona | GPT-5.4 mini | $0.001 / request | $4.00 |
| cometapi | GPT-5.4 mini | $0.60 | $3.60 |
| poe | GPT-5.4 mini | $0.68 | $4.09 |
| jiekou | gpt-5.4-mini | $0.71 | $4.28 |
| edenai | GPT-5.4 mini | $0.75 | $4.50 |
| edenai | GPT-5.4 mini | $0.75 | $4.50 |
| openai | GPT-5.4 Mini | $0.75 | $4.50 |
| 302ai | gpt-5.4-mini | $0.75 | $4.50 |
| abacus | GPT-5.4 Mini | $0.75 | $4.50 |
| aihubmix | GPT 5.4 Mini | $0.75 | $4.50 |
| azure | GPT-5.4 Mini | $0.75 | $4.50 |
| digitalocean | GPT-5.4 mini | $0.75 | $4.50 |
Live data: this table is built from today's prices.
GPT-5.4 nano: every provider
| Provider | Model | Input, $ | Output, $ |
|---|---|---|---|
| openrouter | OpenAI: GPT-5.4 Nano (batch) | $0.10 | $0.62 |
| requesty | GPT-5.4 nano | $0.10 | $0.62 |
| vercel | GPT 5.4 Nano (Batch) | $0.10 | $0.62 |
| cometapi | GPT-5.4 nano | $0.16 | $1.00 |
| poe | GPT-5.4 nano | $0.18 | $1.14 |
| jiekou | gpt-5.4-nano | $0.19 | $1.19 |
| edenai | GPT-5.4 nano | $0.20 | $1.25 |
| edenai | GPT-5.4 nano | $0.20 | $1.25 |
| openai | GPT-5.4 nano | $0.20 | $1.25 |
| 302ai | gpt-5.4-nano | $0.20 | $1.25 |
| abacus | GPT-5.4 Nano | $0.20 | $1.25 |
| aihubmix | GPT 5.4 Nano | $0.20 | $1.25 |
| azure | GPT-5.4 Nano | $0.20 | $1.25 |
| digitalocean | GPT-5.4 nano | $0.20 | $1.25 |
| edenai | GPT-5.4 nano | $0.20 | $1.25 |
Live data: this table is built from today's prices.
How to choose
- Hard reasoning and long documents: GPT-5.5, or GPT-5.4 Pro when quality matters more than a sixfold price.
- Everyday assistant and coding work: GPT-5.4.
- High volume - classification, extraction, routing: GPT-5.4 mini, or nano at about a quarter of mini's price.
- Repeated long prompts: cached input costs a tenth of the normal input price. All OpenAI models in one table: OpenAI at every provider. To estimate a month, use the cost calculator.
All prices in US dollars per 1 million tokens, before taxes, as the providers publish them; every row links to a card with its source and check date.