Rarely has a product release been as anticipated as GPT-5’s, but it appears that it might end up being delayed.
OpenAI CEO Sam Altman had shared a post on X which could indicate that GPT-5, OpenAI’s most powerful AI model, could be delayed. “We have a ton of stuff to launch over the next couple of months–new models, products, features, and more,” Altman said on X. “Please bear with us through some probable hiccups and capacity crunches. Although it may be slightly choppy, we think you’ll really love what we’ve created for you!” he added.
While Altman did not explicitly say that GPT-5 would be delayed, he did speak of “couple of months”, and “probable hiccups” and “capacity crunches”. Altman’s post came at a time when the anticipation for GPT-5 was reaching a crescendo on X, with users expecting the launch to happen at any moment.
It had earlier been reported by Verge that OpenAI would launch GPT-5 in August. This had been accompanied by reports that GPT-5 was being tested on platforms like Hugging Face, and that GPT-5’s launch page on OpenAI’s website had been indexed by Google. Altman, though, seemed to be looking to temper the enthusiasm by not mentioning GPT-5 explicitly, but instead referring to several products. He also mentioned a timeline of a “couple of months”, which could mean that GPT-5’s arrival isn’t imminent.
This wouldn’t be the first time that OpenAI’s had to delay a launch of an anticipated model. OpenAI had said in April that it was excited to release an open-weights model in the “coming months”. In May, Altman had thanked developers for their feedback on the open-weights model, and said they’d ship “something extraordinary”. However in June, Altman had said that the open-weights model would be delayed. “We are going to take a little more time with our open-weights model, i.e. expect it later this summer but not June. Our research team did something unexpected and quite amazing and we think it will be very very worth the wait, but needs a bit longer,” he had said.
It had then been rumoured that OpenAI had decided to push back the date of its own open-weights model following the release of powerful open-source models like Kimi K2. These models had pushed the frontier of what open-source models were capable of, and were likely at par — or even ahead — of what OpenAI was planning to release. As such, it’s possible that OpenAI had delayed its launch to be able to deliver a model that was much superior to any other open models available.
Altman’s latest messaging seems similar to the post he’d shared in May about the open-weights model. Back then, he’d not given a date of when the model would be shipped, but had mentioned it would be “extraordinary”. Today, Altman has again given a vague two-month time-frame, and said that people would “love” what they’d created.
And it’s possible that GPT-5 is delayed for similar reasons as the open-source model. Last month, xAI had released Grok 4, which had gone past OpenAI and Google on most benchmarks. Just yesterday, Google had released a version of the Deep Think model which had won gold at the International Mathematics Olympiad, and had again smashed benchmarks. It’s possible that these models perform roughly at par with what GPT-5 is managing, and OpenAI is looking to push back the launch date to announce a model that’s head and shoulders better than the competition. It remains to be seen when GPT-5 is finally released — it’s possible that all this speculation amounts to nothing, and OpenAI does indeed release GPT-5 in August — but Sam Altman’s latest post seems to be indicating that this might not necessarily be the case.