OpenAI GPT models
OpenAI offers an array of models that showcase varied capabilities and cost considerations. These models span from GPT-4, GPT-3.5, to GPT-3 Legacy, with each having unique attributes and enhancements.
GPT-4
Positioned as the most recent and technologically advanced model in the GPT series, GPT-4 is a large multimodal model equipped to tackle challenging tasks more accurately than any prior OpenAI models. This is attributable to its expansive general knowledge and advanced logical abilities. GPT-4 is primarily optimized for chat, but it is equally competent for traditional completions tasks via the Chat Completions API.
GPT-4 offers different versions, including gpt-4, gpt-4-0613, gpt-4-32k, and gpt-4-32k-0613, each version featuring unique attributes and capabilities. To illustrate, gpt-4-32k maintains similar capabilities to the base gpt-4 model but offers a 4 times longer context length.
Model | Description | Max Tokens |
GPT-4 | More capable than any GPT-3.5 model, able to do more complex tasks, and optimized for chat. | 8,192 |
gpt-4-0613 | Snapshot of gpt-4 from June 13th 2023 with function calling data. | 8,192 |
gpt-4-32k | Same capabilities as the base gpt-4 mode but with 4x the context length. | 32,768 |
gpt-4-32k-0613 | Snapshot of gpt-4-32 from June 13th 2023. | 32,768 |
GPT-3.5
GPT-3.5 models are capable of understanding and generating either natural language or code. The most effective and cost-efficient model in the GPT-3.5 suite is gpt-3.5-turbo. This model is primarily optimized for chat but performs well in traditional completions tasks too.
Model | Description | Max Tokens |
gpt-3.5-turbo | Most capable GPT-3.5 model and optimized for chat at 1/10th the cost of text-davinci-003. | 4,096 |
gpt-3.5-turbo-16k | Same capabilities as the standard gpt-3.5-turbo model but with 4 times the context. | 16,384 |
gpt-3.5-turbo-0613 | Snapshot of gpt-3.5-turbo from June 13th 2023 with function calling data. | 4,096 |
gpt-3.5-turbo-16k-0613 | Snapshot of gpt-3.5-turbo-16k from June 13th 2023. | 16,384 |
Note: Differences between model gpt-3.5-turbo and gpt-3.5-turbo-0301
GPT-3 Legacy
GPT-3 Legacy models are proficient in understanding and generating natural language. These models were succeeded by the more potent GPT-3.5 generation models. However, the original GPT-3 base models, namely davinci, curie, ada, and babbage, are the only models currently available for fine-tuning.
Model | Description | Max Tokens |
text-curie-001 | Very capable, faster and lower cost than Davinci. | 2,049 |
text-babbage-001 | Capable of straightforward tasks, very fast, and lower cost. | 2,049 |
text-ada-001 | Capable of very simple tasks, usually the fastest model in the GPT-3 series, and lowest cost. | 2,049 |
davinci | Most capable GPT-3 model. Can do any task the other models can do, often with higher quality. | 2,049 |
curie | Very capable, but faster and lower cost than Davinci. | 2,049 |
babbage | Capable of straightforward tasks, very fast, and lower cost. | 2,049 |
ada | Capable of very simple tasks, usually the fastest model in the GPT-3 series, and lowest cost. | 2,049 |
import openaiopenai.api_key = "{{SECRET_KEY}}"response = openai.Completion.create(engine="{{Model}}", # You can change this above to "gpt-3.5-turbo" or any other modelprompt="Translate the following English text to French: '{}'",max_tokens=60)print(response.choices[0].text.strip())
Within this code snippet, you have the option to modify the engine parameter to the model of your choice. For instance, gpt-4 can be replaced with gpt-3.5-turbo if you wish to leverage the GPT-3.5 Turbo model.
