OpenAI launched its anticipated replace to GPT-5 on Thursday, boasting that the brand new AI is “probably the most succesful mannequin sequence but for skilled information work.”
“We designed GPT‑5.2 to unlock much more financial worth for individuals; it is higher at creating spreadsheets, constructing shows, writing code, perceiving photographs, understanding lengthy contexts, utilizing instruments, and dealing with complicated, multi-step mission,” the corporate mentioned in a press release.
In a benchmark take a look at known as GDPval, OpenAI mentioned its new AI mannequin can outperform “business professionals at well-specified information work duties spanning 44 occupations.”
“GPT‑5.2 Pondering produced outputs for GDPval duties at >11x the pace and <1% the price of skilled professionals, suggesting that when paired with human oversight, GPT‑5.2 may help with skilled work,” the corporate mentioned.
OpenAI
And in a be aware that’s certain to be a magnet for bankers, OpenAI wrote that in an inner benchmark of junior funding banking analyst spreadsheet modeling duties — “akin to placing collectively a three-statement mannequin for a Fortune 500 firm with correct formatting and citations, or constructing a leveraged buyout mannequin for a take-private” — the brand new mannequin’s rating per process was “9.3% greater than GPT‑5.1’s, rising from 59.1% to 68.4%” on common.
OpenAI mentioned that GPT-5.2 will start rolling out at the moment for paid ChatGPT plans. Paid customers may have entry to GPT-5.1 for 3 months underneath legacy fashions earlier than it’s sunsetted.
“We deploy GPT‑5.2 steadily to maintain ChatGPT as easy and dependable as we are able to,” the corporate mentioned.
The corporate additionally touted its features in agentic coding capacity.
“Even with out the power to do new issues like output polished recordsdata, GPT-5.2 appears like the most important improve we have had in a very long time. Curious to listen to what you suppose!” OpenAI CEO Altman wrote on X.
OpenAI CEO of Purposes Fidji Simo mentioned extra adjustments shall be coming subsequent 12 months as soon as OpenAI rolls out age verification throughout ChatGPT. She mentioned an “grownup mode” for the chatbot will debut within the first quarter of 2026.
The discharge comes simply over per week after Altman declared a “code crimson” in a non-public message to workers, marshaling extra sources to ChatGPT amid rising competitors from Google and different corporations.
“Code reds aren’t unusual,” Somo mentioned throughout an interview on TBPN, including that the hosts may “be the choose” of the outcomes of Altman’s declaration. She mentioned she was “fairly pleased with” GPT-5.2’s advances, though its improvement pre-dated the “code crimson.”
Google has been thought-about by many in tech to be gaining, if not surpassing, OpenAI within the AI race with its current launch of Gemini 3.
The announcement additionally occurred hours after OpenAI brokered a serious cope with Disney, which secured a $1 billion funding and entry to the media large’s profitable and widespread IP.
“It has been a really cool final 10 years; OpenAI has been extra particular to work on than I may have probably imagined,” Altman wrote on X on Thursday.
Leave a Reply