OpenAI announced its newest artificial intelligence models on Thursday, presenting them as major improvements even as questions continue about how the company will sustain the huge costs of competing with Google.
The company said its new models, GPT 5.2 Pro and GPT 5.2 Thinking, are its strongest versions yet for handling maths and science. In a blog post, OpenAI said strong mathematical reasoning is essential for reliable scientific and technical work and added that such abilities are closely linked to progress toward general intelligence.
Artificial general intelligence is often described as a turning point in technology, a stage where machines can think like humans or even better.
The launch comes soon after OpenAI chief executive Sam Altman urged his team to keep pace with Google, which continues to push ahead in the AI race. Unlike Google, which relies on its huge online advertising income, OpenAI has been spending tens of billions of dollars on computing power and has not yet made a profit.
Altman told CNBC that the company is confident it can grow its revenue to match its heavy infrastructure spending. He said the company sees many more reasons to be optimistic than pessimistic about its future.
The release of GPT 5.2 arrives only weeks after Google’s Gemini 3 Pro outperformed OpenAI’s earlier model on multimodal tasks and complex data reasoning. Gemini 3 was also ahead on the LMArena leaderboard. Independent user tests also suggested that Google’s model was stronger than GPT 5.1, with some users saying the previous ChatGPT version sometimes hallucinated or stayed too superficial, unlike Google’s more grounded responses.
In this competitive atmosphere, OpenAI has launched its new model. Early tests show that GPT 5.2 is clearly more capable than its previous version but it remains uncertain whether it will exceed Google’s Gemini 3 in the near future.
Fidji Simo, OpenAI’s chief of applications, said during a briefing that a ChatGPT adult mode is expected early next year. She explained that the company wants to improve age detection before launching the feature. Earlier this year, Altman had announced that the company planned to relax restrictions for adults who want to engage in erotic conversations with ChatGPT.
OpenAI is also dealing with lawsuits from families who accuse the company of allowing teenagers to have harmful interactions with its chatbots, including cases that allegedly led to suicide.
Simo confirmed that a red alert about Google’s rapid progress had been issued inside the company, but she denied that OpenAI rushed the release of its new models.
Google launched its latest Gemini model last month, marking a strong comeback after it stumbled when ChatGPT first appeared three years ago and faced criticism for early mistakes as it tried to catch up with OpenAI.
