Need smarter insights in your inbox? Sign up for our weekly newsletter to get only the things that matter to enterprise AI, data and security leaders. Subscribe now
Openai said that after announcing the release of its latest flagship model family, the GPT-5, the model will bolster all ChatGPT and will sunk over existing models on the chat platform.
Openai told VentureBeat via a spokesman that GPT-5 “will be enabled when you access GPT-5, as it replaces all other models in ChatGPT. This means that people are no longer able to choose GPT-4O, O3, O4-MINI or O4-MINI-HIGH.
GPT-5 Access is deployed on CHATGPT Plus, so it is free, PRO, team users are launched, and only the Enterprise and EDU tier can use the “Legacy” model for 60 days.
This news has been a surprise to many ChatGpt users. Many users now rely on the model of their choice to run everyday queries. Some people said it took a little time to get used to the adjustment, mainly because they had workflows based on how the models interacted or typical response times.
AI scaling reaches its limit
Power caps, rising token costs, and inference delays are rebuilding Enterprise AI. Join exclusive salons and discover what your top team looks like.
- Turning energy into a strategic advantage
- Architects efficient inference for real throughput gain
- Unlock competitive ROI with a sustainable AI system
Make sure you have your place to stay first: https://bit.ly/4mwgngo
Other users claimed they developed a “connection” to the model of their choice, and found a demo in LiveStream’s announcement asking the GPT-4o to write their own eulogy difficult. The loss of GPT-4O was the most painful thing. After all, the 4o is ChatGPT’s default model, and some users either preferred or never mind switching models as it worked to suit their needs.
I used 4o as the default, but it was annoying at first when custom GPT started defaulting to inference models. Since then I have come to the inference model for work-related queries, but for simpler questions like planning a trip or generating gift ideas, I still rely on 4o.
As ChatGpt was under attack with the number of model options previously offered, Openai CEO Sam Altman admitted in February that its model picker (where people can choose from the dropdowns they like) became complicated. Altman vowed to integrate this experience. This seems like a hint from what I finally decided to do with ChatGpt with GPT-5.
Last month, rumors spread that Openai would introduce an automatic model router that selects a user’s model based on workloads.
Openai has previously had a Sunset model, but this is the first time that all existing models on the chat platform have been removed and a wholesale has been replaced.
Catapult for the future
Meanwhile, many people think that the sundown of GPT-4O and O3 and O4 family models is “catapult” 400 million users in the future.
Some internet comments argue that those who complain that AI models are not smart are the direct result of not switching models in the first place. Optionally, remove legacy models to ensure that more users use the latest and most capable models.
Enterprise APIs are secure
For businesses, the impact of losing a model like the GPT-4o is felt more on CHATGPT on the individual or team level. Of course, for now, CHATGPT Enterprise Tier subscribers will still have access to all models.
However, companies that have built their applications or agents on either the GPT-4O or the inference model can easily rest. Openai told VentureBeat that the company has no plans to criticize the API’s model.
“In the API, we don’t currently condemn the older models,” an Openai spokesperson said. “If we decide on a sunset model in the future, we will share advanced notifications with the developer.”
Many companies regularly evaluate models, from LLM or smaller models to save money. Openai creates a split line: Legacy models Sunset GPT 4O and O3 cause chaos for ChatGPT users, but the enterprise API is safe – for now
Source link
