GPT AI Power exclusively supports OpenAI GPT-3, GPT-3.5 and GPT-4 models.
Please be aware that GPT-4 is currently in limited beta, which means that access to the GPT-4 API from OpenAI is available only through a waiting list and is not open to everyone yet. You can sign up for the waiting list at here. If you dont have GPT-4 API access and if you select GPT-4 as your model then you will get “The model: gpt-4 does not exist” or “The model: gpt-4-32k does not exist” error.
Supported GPT-4 models:
Model | Description | Max Tokens | Training Data |
---|---|---|---|
gpt-4 | More capable than any GPT-3.5 model, able to do more complex tasks, and optimized for chat. Still in beta. | 8,192 | Up to Sep 2021 |
gpt-4-32k | Same capabilities as the base gpt-4 mode but with 4x the context length. | 32,768 | Up to Sep 2021 |
Supported GPT-3.5 models:
Model | Description | Max Tokens | Training Data |
---|---|---|---|
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 | Up to Sep 2021 |
text-davinci-003 | Can do any language task with better quality, longer output, and consistent instruction-following. | 4,000 | Up to Jun 2021 |
text-davinci-002 | Similar capabilities to text-davinci-003 but trained with supervised fine-tuning instead of reinforcement learning. | 4,000 | Up to Jun 2021 |
Supported GPT-3 models:
Model | Description | Max Tokens | Training Data |
---|---|---|---|
text-curie-001 | Very capable, but faster and lower cost than Davinci. | 2,048 | Up to Oct 2019 |
text-babbage-001 | Capable of straightforward tasks, very fast, and lower cost. | 2,048 | Up to Oct 2019 |
text-ada-001 | Capable of very simple tasks, usually the fastest model in the GPT-3 series, and lowest cost. | 2,048 | Up to Oct 2019 |
OpenAI recommends using Davinci and Turbo while experimenting since they will yield the best results.
Model | Good at |
---|---|
Turbo | Conversation and text generation |
Davinci | Complex intent, cause and effect, summarization for audience |
Curie | Language translation, complex classification, text sentiment, summarization |
Babbage | Moderate classification, semantic search classification |
Ada | Parsing text, simple classification, address correction, keywords |
Selecting a Model
You can change the model used by the plugin by selecting it from the Model dropdown in the AI Engine tab.
You may notice a “Sync” button located next to the model dropdown. This button is used to synchronize the plugin with the most recent version of the OpenAI models. So if OpenAI releases a new model, you can use this button to update the plugin to use the new model.
If your organization has fine-tuned any models on specific datasets, you can easily find them under the Model dropdown.