Asger Risom
OpenAI said the results suggest the model can meaningfully assist professional workflows when paired with human oversight, rather than fully replacing workers.
OpenAI is positioning its latest model release as a step change in how artificial intelligence supports white-collar work, saying GPT-5.2 delivers its strongest performance yet on tasks traditionally handled by trained professionals.
The update arrives amid intensifying competition with rivals such as Google, underscoring how quickly the frontier of enterprise-grade AI is advancing.
A productivity leap
In announcing GPT-5.2, OpenAI said the model is designed to generate “even more economic value” by improving its ability to handle complex, multi-step projects. The company highlighted gains in spreadsheet creation, presentation building, software development, image understanding and long-context reasoning.
According to OpenAI, GPT-5.2 represents the company’s most capable model series to date for what it describes as professional knowledge work.
Benchmark claims
To support that claim, OpenAI pointed to results from an internal benchmark called GDPval, which evaluates well-defined knowledge tasks across 44 occupations. The company said GPT-5.2 outperformed industry professionals on these tasks while completing them more than 11 times faster and at less than 1% of the cost.
OpenAI said the results suggest the model can meaningfully assist professional workflows when paired with human oversight, rather than fully replacing workers.
The company also highlighted improvements in financial modeling. In tests simulating junior investment banking analyst work — such as building three-statement models for Fortune 500 companies or constructing leveraged buyout scenarios — GPT-5.2’s average score rose to 68.4%, up from 59.1% in the prior GPT-5.1 release.
Rolling out to users
GPT-5.2 began rolling out Thursday to paid ChatGPT subscribers. OpenAI said GPT-5.1 will remain available as a legacy option for three months before being retired.
The company said the gradual rollout is intended to maintain reliability as usage scales.
Competitive pressure
The release follows reports that OpenAI CEO Sam Altman recently declared a “code red” internally, reallocating resources toward ChatGPT as competition intensifies. Google’s Gemini 3 has been widely viewed in the tech industry as narrowing — or in some cases overtaking — OpenAI’s lead.
Altman wrote on X that GPT-5.2 feels like the biggest upgrade the company has delivered in some time, even without entirely new output formats.
OpenAI’s head of applications, Fidji Simo, said further changes are planned for 2026, including age verification for ChatGPT and the introduction of an “adult mode” early next year.
Momentum beyond models
The GPT-5.2 announcement comes the same day OpenAI finalized a major partnership with Disney, which includes a $1 billion investment and licensing access to some of the media company’s most valuable intellectual property.
Together, the developments signal OpenAI’s effort to strengthen both its technical lead and its commercial footing as AI tools move deeper into professional and creative industries.
Sources: OpenAI, X.com
