Home / OpenAI API pricing
OpenAI API pricing
What OpenAI charges for its GPT models through the API, per million tokens, and what a month of your own use would cost.
Checked October 13, 2026 against OpenAI’s pricing page
Your monthly OpenAI cost
Models
- Turn on JavaScript to see the costs. The price tables below work without it.
Current GPT-6 models
| Model | Input | Cached input | Cache write | Output | Batch in / out | Good to know |
|---|---|---|---|---|---|---|
| GPT-6 Astragpt-6-astra | $10.00 | $1.00 | $12.50 | $50.00 | $5.00 / $25.00 | Over 272,000 tokens in a prompt: $20.00 in, $2.00 cached, $75.00 out1,050,000-token context; up to 922,000 tokens in |
| GPT-6.1 Solgpt-6.1-sol | $2.00 | $0.10 | $2.50 | $10.00 | $1.00 / $5.00 | Over 272,000 tokens in a prompt: $4.00 in, $0.20 cached, $15.00 out1,050,000-token context; up to 922,000 tokens in |
| GPT-6 Lunagpt-6-luna | $0.10 | $0.01 | $0.13 | $0.50 | $0.05 / $0.25 | Over 272,000 tokens in a prompt: $0.20 in, $0.02 cached, $0.75 out1,050,000-token context; up to 922,000 tokens in |
US dollars per 1 million tokens, standard tier, from OpenAI’s pricing page, checked Oct 13, 2026.
Older models still listed
| Model | Input | Cached input | Cache write | Output | Batch in / out | Good to know |
|---|---|---|---|---|---|---|
| GPT-6 Solgpt-6-sol | $2.00 | $0.20 | $2.50 | $10.00 | $1.00 / $5.00 | Over 272,000 tokens in a prompt: $4.00 in, $0.40 cached, $15.00 out |
| GPT-5.6 Solgpt-5.6-sol | $4.00 | $0.40 | $5.00 | $20.00 | $2.00 / $10.00 | Over 272,000 tokens in a prompt: $8.00 in, $0.80 cached, $30.00 out |
| GPT-5.6 Terragpt-5.6-terra | $2.00 | $0.20 | $2.50 | $12.00 | $1.00 / $6.00 | Over 272,000 tokens in a prompt: $4.00 in, $0.40 cached, $18.00 out |
| GPT-5.6 Lunagpt-5.6-luna | $0.20 | $0.02 | $0.25 | $1.20 | $0.10 / $0.60 | Over 272,000 tokens in a prompt: $0.40 in, $0.04 cached, $1.80 out |
| GPT-5.5gpt-5.5 | $5.00 | $0.50 | — | $30.00 | $2.50 / $15.00 | Over 272,000 tokens in a prompt: $10.00 in, $1.00 cached, $45.00 out |
| GPT-5.4gpt-5.4 | $2.50 | $0.25 | — | $15.00 | $1.25 / $7.50 | Over 272,000 tokens in a prompt: $5.00 in, $0.50 cached, $22.50 out |
| GPT-5.4 minigpt-5.4-mini | $0.75 | $0.075 | — | $4.50 | $0.38 / $2.25 | |
| GPT-5.4 nanogpt-5.4-nano | $0.20 | $0.02 | — | $1.25 | $0.10 / $0.63 | |
| GPT-5 minigpt-5-mini | $0.25 | $0.025 | — | $2.00 | $0.13 / $1.00 | |
| GPT-5 nanogpt-5-nano | $0.05 | $0.005 | — | $0.40 | $0.025 / $0.20 | |
| GPT-4.1gpt-4.1 | $2.00 | $0.50 | — | $8.00 | $1.00 / $4.00 | |
| GPT-4.1 minigpt-4.1-mini | $0.40 | $0.10 | — | $1.60 | $0.20 / $0.80 | |
| GPT-4o minigpt-4o-mini | $0.15 | $0.075 | — | $0.60 | $0.075 / $0.30 |
US dollars per 1 million tokens, standard tier, from OpenAI’s pricing page, checked Oct 13, 2026.
OpenAI lists more older models on its pricing page. A dash means OpenAI shows no price for that item.
What changes the bill
Long context
OpenAI prices GPT-6, GPT-5.6, GPT-5.5 and GPT-5.4 in two bands: up to 272,000 input tokens, and over 272,000. Above the line, input costs twice as much and output 1.5 times as much. GPT-6.1 Sol, for example, goes from $2.00 in and $10.00 out to $4.00 and $15.00.
Cached input and cache writes
Repeated input read from cache costs a tenth of the input price on GPT-6 Astra and Luna, and a twentieth on GPT-6.1 Sol. The GPT-6 and GPT-5.6 models also list a cache-write price, 1.25 times the input price.
Processing tiers
- Batch: half the standard price, for requests that can wait. The Flex tier is also half price on the GPT-6 and GPT-5 models.
- Fast: twice the standard price on the GPT-6 models. OpenAI renamed Priority processing to Fast mode on 30 July 2026.
- Ultrafast: six times the standard price, listed for GPT-6 Astra ($60 in, $300 out per million) and GPT-6.1 Sol ($12 in, $60 out).
Other charges
- Regional data-residency endpoints cost 10% more for models released on or after 5 March 2026. FedRAMP endpoints also cost 10% more.
- Web search costs $10 per 1,000 calls, plus the search content tokens at the model’s rates.
- GPT-5.6 Sol is on a promotional price that OpenAI says runs at least through 21 November 2026.