💰 Pricing

Our API service billing is based on the official method, which makes it very convenient for you to track and manage your expenses.

Recharge Ratio

1 US dollar = 1 US dollar of OpenAI official credits

When you recharge, we will credit the corresponding amount to your account based on the amount you recharged (in US dollars).

Your credits will be consumed according to the model you use.

Model Pricing

Discover amazing savings and get the best of both worlds with GPT-3.5-turbo and GPT-4 models. We offer GPT-3.5-turbo at one-third of the official price and GPT-4 at 1.1x the official price. For example:

  1. With GPT-3.5-turbo, spend $20 on our platform and get 30 million tokens, compared to 10 million tokens at OpenAI's official pricing.
  1. For GPT-4, one million tokens cost $30 at OpenAI's official site, while the same amount of tokens on our platform only requires $33.

Friendly Reminder

If you are a corporate or team user, we can even offer you more discounts (including the GPT-4 model) based on the above. Please contact our Telegram customer service immediately to start enjoying your exclusive benefits.

The best part is: besides GPT-4, other supported models are officially defined prices open in new window one-third of the cost.

GPT-3

Unit: USD

Model $/1K tokens $/10M tokens
gpt-3.5-turbo 0.002*(1/3) 20*(1/3)

For example:

  • 1,000 tokens cost 0.002*(1/3) US dollars
  • 10M tokens cost 20*(1/3) US dollars

GPT-4

Unit: USD

TIP

gpt-4 currently supports the 8K context version. Passing in gpt-4 defaults to 8K, and gpt-4-32k requires customer service to manually enable it for you.

Model $/1K tokens $/1M tokens
8K prompt 0.03*1.1 30*1.1
8K completion 0.6*1.1 60*1.1

For example:

  • 1,000 tokens cost 0.03*1.1 US dollars
  • 1M tokens cost 30*1.1 US dollars

GPT

Unit: USD

Model $/1K tokens $/1M tokens
Ada 0.0004*(1/3) 0.4*(1/3)
Babbage 0.0005*(1/3) 0.5*(1/3)
Curie 0.0020*(1/3) 2*(1/3)
Davinci 0.02*(1/3) 20*(1/3)

For example (Ada):

  • 1,000 tokens cost 0.0004*(1/3) US dollars
  • 1M tokens cost 0.4*(1/3) US dollars

DALL·E Image

Unit: USD

Size $/image
1024×1024 0.03*(1/3)
512×512 0.027*(1/3)
256×256 0.024*(1/3)

Whisper Voice

Model Usage
Whisper 0.006 USD*(1/3) / minute (rounded to the nearest second)

API Calls and Charges

When you call our API service, the system will calculate the cost in real-time based on the model you use and the number of tokens consumed, and deduct it from your account balance. If your balance is insufficient, we will prompt you to recharge.

GPT-4 Pricing

Due to the high cost of gpt-4 and the low pricing of other models on our platform, we had to adjust the calculation multiplier for gpt-4.

Now, the gpt-4 model is temporarily set at 1.1 times the official model price, and currently supports the 8Kand32k version.

In summary, calling other models will be charged normally, while calling gpt-4 will be charged at 1.1x the official rate.

Additionally, based on our experience processing billions of user tokens, you can delegate core project tasks to GPT-4, and use GPT-3.5 for the remaining, less complex tasks. This approach will still result in lower overall costs compared to the official rates.