OpenAI’s GPT-4 Turbo receives stealth update for 2023

OpenAI’s most advanced commercial AI model, GTP-4 Turbo, has been quietly updated to a data training set as recently as December 2023 without any public announcement from the company.

OpenAI’s GPT-4 Turbo receives stealth update for 2023

The artificial intelligence (AI) developer OpenAI has quietly updated the training data set for its most advanced model, GPT-4 Turbo, to be relevant from December 2023, according to documentation on its website. 

Turbo’s latest update, which was not widely announced by the developer, means that it is now equipped with data relevant up until December 2023, making it the most up-to-date model offered by OpenAI. By comparison, its free model, GPT 3.5, is only equipped with data relevant up until January 2022.

According to the site, the update was implemented with the intention of reducing what developers call “laziness” in a model, in which it doesn’t complete tasks asked of it.

At its DevDay in November, OpenAI revealed new models and developer products, during which it said GPT-4 Turbo would be knowledgeable of current events up until April 2023. Therefore, the latest update equips the model with eight more months of data and information.

In its most recent update on Jan. 25, there was no mention of the update to the training data set for GPT-4 Turbo.

Related: Groq AI model goes viral and rivals ChatGPT, challenges Elon Musk’s Grok

While website information shows updated data until December 2023, the OpenAI developer forum began to buzz with questions of legitimacy.

Some users were observing that their preview doesn’t seem to be updated to December 2023; instead, they were returning test prompts with answers like “as of April 2023…,” which was the previous update timeframe.

Source: OpenAI developer forum

However, responses from other users have revealed positive results, with models responding that their information includes data updated from December 2023. 

Source: OpenAI developer forum

This latest update to OpenAI’s most advanced model comes just days after it unveiled its advanced text-to-video model, Sora, which left users on social media in awe of its realistic outputs. 

OpenAI claims Sora can generate movie-like scenes containing multiple characters, specific motions and subjects, and background details in up to 1080p resolution. However, the developer said the model is still not ready for public release.

On Feb. 17, a report in The New York Times said OpenAI’s valuation had skyrocketed to $80 billion in its latest deal. The company is currently in talks with investors around the world, as well as international governments, to raise funds to develop AI chips in-house.

Magazine: ChatGPT trigger happy with nukes, SEGA’s 80s AI, TAO up 90%: AI Eye

Related Articles

Responses