OpenAI reveals an updated GPT-4o model – but can’t quite explain how it’s better

There is a new model of OpenAI’s GPT-4o mannequin on the town. However what it might probably exactly do appears to be a thriller, even to OpenAI. In an X submit on Monday, the corporate spilled the beans, saying: “there is a new GPT-4o mannequin out in ChatGPT since final week. Hope you all are having fun with it and test it out if you have not! we predict you will prefer it.”

In any other case, OpenAI was mum about what enhancements this new mannequin presents. In updates to its X submit, the corporate stated that the brand new GPT-4o mannequin is accessible for paid subscribers in addition to these on the free tier (with a message cap). But it surely’s not GPT-4o-2024-08-06, which was additionally launched final week and is now operating on Microsoft Azure.

Some ChatGPT customers chimed in earlier than Monday’s announcement, claiming they seen a distinction within the chatbot’s dealing with of requests and duties. In accordance with VentureBeat, a number of folks felt that GPT-4o was behaving in another way and higher than previously. Others stated that GPT-4o’s native picture technology abilities by means of ChatGPT appeared to be kicking in. A couple of stated that the improve improved multi-step reasoning.

In a single  X submit, an account named @misaligned_agi stated, “Wow, GPT-4o now makes use of multi-step reasoning. It is spectacular to see this in motion. Seems the replace wasn’t a brand new mannequin however a brand new technique.”

With multi-step reasoning, an AI breaks down advanced issues and questions right into a smaller collection of sequential steps, tackling every step individually, after which comes up with the response. The most effective instance is a math drawback that requires a number of calculations. The AI solves every equation to reach on the general reply.

Nevertheless, a spokesperson for OpenAI informed me that the hypothesis about multi-step reasoning missed the mark.

After a lot theorizing amongst ChatGPT customers, OpenAI lastly shed some mild in regards to the replace, now referred to as ChatGPT-4o-latest. The one factor is that the corporate’s rationalization continues to be imprecise.

“Bug fixes and efficiency enhancements … we have launched an replace to GPT-4o that we have discovered, by means of experiment outcomes and qualitative suggestions, ChatGPT customers are likely to choose,” OpenAI stated in its newest launch notes on Tuesday. “It isn’t a brand new frontier-class mannequin. Though we would wish to inform you precisely how the mannequin responses are totally different, determining granularly benchmark and talk mannequin habits enhancements is an ongoing space of analysis in itself (which we’re engaged on!).”

This means that OpenAI conjured up a brand new and improved mannequin however does not actually know the way or why it is higher. Hmm, OK. Additional particulars within the launch notes nonetheless did not reply the query.

“Typically we are able to level to new capabilities and particular enhancements — and we’ll attempt our greatest to speak that each time doable,” OpenAI added in its notes. “Within the meantime, our crew is consistently iterating on the mannequin by including good information, eradicating dangerous information, and experimenting with new analysis strategies based mostly on person suggestions, offline evaluations, and extra. That is the case with this mannequin replace.”

Right here, it appears like OpenAI is ready for customers to outline the brand new mannequin so that everybody can work out what it really does. In different phrases, OpenAI says to its customers, “You inform me, after which we’ll each know.”

On its ChatGPT fashions web page, the corporate offered a couple of specifics on ChatGPT-4o-latest. Described as a dynamic mannequin repeatedly up to date to the present model of GPT-4o, it is meant for analysis and analysis.

Skilled on information as much as October 2023, this newest mannequin can deal with 128,000 tokens, or 96,000 phrases, in a single dialog, the identical quantity as its predecessors. Nevertheless, it might probably output as much as 16,384 tokens, or 12,288 phrases, the identical as GPT-4o-mini, however with an enchancment of over 4,096 tokens within the authentic GPT-4o mannequin.

No matter new mannequin or technique OpenAI has added to GPT-4o, the outcomes actually appear well worth the effort. The most recent model landed on the prime of the pack in testing at Chatbot Area, a website that pits one AI chatbot mannequin towards one other.

Listed beneath “anonymous-chatbot,” ChatGPT-4o-latest earned a rating of 1315 based mostly on greater than 11,000 group votes, serving to OpenAI reclaim the highest spot from Google’s Gemini 1.5. Primarily based on its efficiency, the brand new mannequin confirmed a notable enchancment in such technical domains as coding, following directions, and onerous prompts.

If you wish to see for your self, taking ChatGPT-4o-latest for a spin your self is easy sufficient. The brand new abilities are already baked into the model of GPT-4o accessible with the ChatGPT web site and cell apps (in addition to the API). ChatGPT Plus subscribers ought to ensure that the mannequin is about to GPT-4o, whereas free customers can use the usual ChatGPT.

Attempt asking extra advanced and nuanced questions and see how the AI fares, particularly in contrast with its previous efficiency. Then, perhaps collectively, we’ll work out what this new mannequin really does.