OpenAI
, the company behind ChatGPT, has announced that it is releasing new
language models
and updating the existing models so that they can undertake tasks more thoroughly than the previous preview model.
The new models include:
- Two new embedding models
- An updated GPT-4 Turbo preview model
- An updated GPT-3.5 Turbo model
- An updated text moderation model
What are the new embedding models
OpenAI calls the two new embedding models as text-embedding-3-small model which is the smaller of the two models; and the more powerful text-embedding-3-large model.
OpenAI claims that the text-embedding-3-small “provides a significant upgrade over its predecessor” that was released in December 2022. It provides “stronger performance”, is “more efficient” and will be available at a reduced price (by 5X) compared to the previous generation model.
Updated GPT-3.5 Turbo model
The free version of ChatGPT uses the GPT-3.5 Turbo model and the company will be rolling out various improvements, including higher accuracy at responding in requested formats and a fix for a bug which caused a text encoding issue for non-English language function calls. The company is also reducing the price of the model.
Updated GPT-4 Turbo preview
The company announced GPT-4 Turbo as the more powerful model than the GPT-4 late last year and OpenAI said that over 70% of requests from GPT-4 API customers have transitioned to GPT-4 Turbo since its release.
The company released an updated GPT-4 Turbo preview model, which is said to complete tasks like
code generation
more thoroughly than the previous preview model. It is “intended to reduce cases of ‘laziness’ where the model doesn’t complete a task. The new model also includes the fix for the bug impacting non-English UTF-8 generations.”
English (US) ·