- Sam Altman says GPT-5.4 is his “favorite talking model” and a step forward for ChatGPT’s personality
- OpenAI admits that the model still has three notable weaknesses, including design taste, real-world context, and task completion.
- The feedback suggests that OpenAI is now focusing as much on how ChatGPT feels when using it as it does on its power.
In a recent post on while admitting that previous fifth-generation models had not been perfect when it came to human interaction. “We’ve lost our way in terms of model personality for a while, so it feels really good to be moving in the right direction,” she added.
His use of the phrase “for a while” appears to reference the ChatGPT personality debate, which has dogged OpenAI since it retired the popular ChatGPT-4o model in November last year. Many users felt that the 4o had a much better personality than later fifth-generation GPT models.
There have been campaigns to bring back the 4o model since it was retired on February 13, 2026, despite access to fifth-generation GPT models being widely available.
Article continues below.
OpenAI has recently launched new GPT-5.3 Instant and GPT-5.4 Thinking models, and it is this 5.4 model that Altman believes has the best personality, as well as being excellent at coding.
GPT-5.4 is great for coding, knowledge work, using computers, etc., and it’s nice to see how much people enjoy it. But he’s also my favorite model to talk to! We’ve lost our way on model personality for a while, so it feels really good to be moving in the right direction.March 7, 2026
The best model in the world.
However, Shumer went on to list three weaknesses. Specifically:
- “The frontend flavor is far behind Opus 4.6 and Gemini 3.1 Pro. Why is it so hard to fix? @OpenAI once you fix this there is literally no reason for me to use any other model. Please, please do it!
- You can still miss the obvious real-world context. For example, I asked him to plan an itinerary for a trip. At first glance, it seemed perfect, but it didn’t take into account that I was choosing locations that would be crowded with spring break students, so I had to re-run the message from scratch with more context.
- When testing it within OpenClaw, it stopped before finishing the tasks. I guess this will be fixed quickly, but it’s still worth noting.”
In response, Altman enthusiastically posted: “We will be able to fix these three things!”
You might be wondering what “frontend taste” means here. We think Shumer refers to AI’s sense of style and aesthetics when generating user interfaces for things like web applications, so this relates to coding.
While praising GPT-5.4’s coding capabilities, it appears that the user interfaces it produces are less polished, modern, or visually appealing than those generated by rivals like Claude or Gemini.
The other two reviews seem simpler, although if you’re wondering what OpenClaw is, it’s an open source system designed to automate AI workloads on a large number of Macs, particularly groups of Mac minis. It can be used to combine multiple Macs to run or train AI models.
The popularity of OpenClaw is blamed for shortages in the Mac mini supply chain.
More than anything, the exchange hints that OpenAI is now focused on how ChatGPT feels to use, not just how well it works on paper. This is significant, because many of the complaints about the fifth-generation GPT models were not really about capability, but rather personality, with people criticizing them for seeming robotic, flat, or too sycophantic.
Altman’s posts suggest that OpenAI knows that getting the tone right is no longer a secondary issue, but a key part of creating a GPT that people actually want to talk to.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
