Regardless of OpenAI persevering with to steer the AI race, the corporate hasn’t slowed down its tempo of growth, persevering with to launch modifications, upgrades, and even new know-how. The most recent wave of updates consists of extremely anticipated adjustments, together with GPT-4 Turbo efficiency enhancements and decrease API prices.
On Thursday, OpenAI unveiled a sequence of updates, together with an up to date GPT-4 Turbo preview. Regardless of being the OpenAI’s most superior AI mannequin, GPT-4 Turbo has been the article of many complaints about its efficiency, together with claims that it was acting lazy.
Additionally: The best AI image generators of 2024: Tested and reviewed
OpenAI has addressed that challenge by releasing an up to date GPT-4 Turbo preview mannequin, gpt-4-0125-preview, which may higher full duties resembling code era in comparison with the earlier mannequin. OpenAI mentioned it might probably even “cut back circumstances of ‘laziness’ the place the mannequin would not full a job.”
The corporate additionally shared a standing replace on the extremely anticipated GPT-4 Turbo with Imaginative and prescient, which shall be OpenAI’s most superior mannequin but, sharing that it plans to launch the mannequin for common availability within the coming months.
Subsequent week, OpenAI additionally will introduce a brand new GPT-3.5 Turbo mannequin, gpt-3.5-turbo-0125, which can sport “numerous enhancements,” in line with OpenAI, together with larger accuracy at responding in requested codecs and a repair for text-encoding bugs for non-English language features calls.
The costs for GPT-3.5 Turbo can even drop for the third time up to now 12 months, with enter costs for the brand new mannequin being lowered by 50% to $0.0005 per 1,000 tokens, whereas output costs are lowered by 25% to $0.0015 per 1,000 tokens.
Additionally: I tried Google’s new Art Selfie 2 and could not stop laughing
Tokens are used because the forex for an OpenAI’s API to carry out a job. The quantity of tokens wanted to finish a job will depend on the size of the enter and output and may usually rack up shortly if you’re utilizing ChatGPT for bigger our bodies of textual content. The worth drop ought to assist offset the price of utilizing ChatGPT for these functions.
The final replace went to the free Moderation API, which builders can use to establish probably dangerous textual content. OpenAI launched its most sturdy moderation mannequin, text-moderation-007.
When it comes to new fashions, OpenAI launched two new embedding fashions: the text-embedding-3-small mannequin and the text-embedding-3-large mannequin. Embedding fashions signify textual content, pure language, or code as vectors, which helps machine studying fashions perceive the relationships between content material.