Our mission is to provide accurate technology knowledge to individuals worldwide, and our website is not monetized.

What is OpenAI’s ChatGPT Turbo? GPT-4 Turbo vs GPT-4

OpenAI announced GPT-4 Turbo, an upgraded version of its GPT-4 large-scale language model, at the OpenAI DevDay conference for developers on November 6, 2023. GPT-4 Turbo will first be offered to users of ChatGPT Plus (paid version) and then to general users.

There are four great things about GPT 4 Turbo, If you have been familiar with GPT 4 via the API, you will be able to do more than ever before!

  • Increased number of characters that can be entered
  • Multimodal on the API
  • Response optimization
  • Reduced usage fees
TOC

Difference between GPT-4 Turbo and GPT-4

Which aspects of GPT-4 Turbo have been updated compared to GPT-4?

スクロールできます
GPT-4 TurboGPT-4
number of charactersUp to 128,000 tokensUp to 32,768 tokens
learning dataThrough April 2023Through September 2021
Fee (input)$0.01$0.03
Fee (output)$0.03$0.06
Data that can be entered via APIText + Imagetext
Data that can be output via APIText + images + soundtext

Increased Character Count: Support for up to 128,000 Tokens The standard GPT-4 model has a context window capped at 8,000 tokens, with a special extension to 32,000 tokens for certain developers. GPT-4 Turbo significantly expands this limit to 128,000 tokens, equivalent to about 300 book pages. This model is recognized for its heightened accuracy and its ability to analyze and summarize a broad range of documents.

Prompts: Capacity for Longer Text Inputs GPT-4 Turbo’s enhanced capability allows it to handle contexts of up to 128,000 tokens, accommodating longer and more complex prompts.

Training Data: Updated Through April 2023 Trained with data up to April 2023, GPT-4 Turbo stands as OpenAI’s most advanced large-scale language model. OpenAI also plans regular updates to GPT-4 Turbo’s database.

Rates: Significant Reduction in Costs GPT-4 Turbo introduces a cost reduction: input tokens are now priced at $0.01, 1/3 the cost of GPT-4’s, while output tokens are $0.03, half of GPT-4’s rate.

Conversely, GPT-3.5 Turbo’s input token cost is $0.001, which is 1/3 of the rate for the older 16K model, and output tokens are priced at $0.002, half of the 16K model’s rate. Developers using GPT-3.5 Turbo 4K will see a 33% reduction in input token cost to $0.001. These reduced rates are specific to the new GPT-3.5 Turbo model.

For the fine-tuned GPT-3.5 Turbo 4K model, input tokens cost $0.003 (1/4 the standard rate), and output tokens are $0.006 (1/2.7 the standard rate). The updated GPT-3.5 Turbo model, including fine-tuning, supports a 16K context at the same pricing as the 4K model. These revised rates also apply to the fine-tuned gpt-3.5-turbo-0613 model.

スクロールできます
Older modelsNew models
GPT-4 TurboGPT-4 8K Input: $0.03 Output: $0.06 GPT-4 32K Input: $0.06 Output: $0.12GPT-4 Turbo 128K Input: $0.01 Output: $0.03
GPT-3.5 TurboGPT-3.5 Turbo 4K Input: $0.0015 Output: $0.002 GPT-3.5 Turbo 16K Input: $0.003 Output: $0.004GPT-3.5 Turbo 16K Input: $0.001 Output: $0.002
GPT-3.5 Turbo fine-tuningGPT-3.5 Turbo 4K fine-tuning Training: $0.008 Input: $0.012 Output: $0.016GPT-3.5 Turbo 4K and 16K fine-tuning Training: $0.008 Input: $0.003 Output: $0.006
https://openai.com/pricing

GPT-4 Turbo

ModelInputOutput
gpt-4-1106-preview$0.01 / 1K tokens$0.03 / 1K tokens
gpt-4-1106-vision-preview$0.01 / 1K tokens$0.03 / 1K tokens

GPT-4

ModelInputOutput
gpt-4$0.03 / 1K tokens$0.06 / 1K tokens
gpt-4-32k$0.06 / 1K tokens$0.12 / 1K tokens

GPT-3.5 Turbo

ModelInputOutput
gpt-3.5-turbo-1106$0.0010 / 1K tokens$0.0020 / 1K tokens
gpt-3.5-turbo-instruct$0.0015 / 1K tokens$0.0020 / 1K tokens

Assistants API

ToolInput
Code interpreter$0.03 / session (free until 11/17/2023)
Retrieval$0.20 / GB / assistant / day (free until 11/17/2023)

Fine-tuning models

ModelTrainingInput usageOutput usage
gpt-3.5-turbo$0.0080 / 1K tokens$0.0030 / 1K tokens$0.0060 / 1K tokens
davinci-002$0.0060 / 1K tokens$0.0120 / 1K tokens$0.0120 / 1K tokens
babbage-002$0.0004 / 1K tokens$0.0016 / 1K tokens$0.0016 / 1K tokens

Embedding models

ModelUsage
ada v2$0.0001 / 1K tokens

Base models

ModelUsage
davinci-002$0.0020 / 1K tokens
babbage-002$0.0004 / 1K tokens

For more detail, please visit New models and developer products announced at DevDay

Let's share this post !
  • Copied the URL !
  • Copied the URL !

Comments

To comment

TOC