ChatGPT maker takes on Google's Gemini 3 with latest AI model GPT-5.2

As the competition in the realm of artificial intelligence (AI) is intensifying, OpenAI has launched its latest AI model, GPT-5.2, claiming it as its most advanced model yet.

Designed for both developers and ordinary users, GPT-5.2 comes in response to Google's Gemini 3, which is currently in the lead in many benchmarks.

GPT-5.2's debut follows a “code red” memo from OpenAI CEO Sam Altman, highlighting a decline in ChatGPT traffic and the need to refocus on improving user experience.

The new GPT-5.2 model is available to ChatGPT paid users and developers via API in three variants: Instant, for routine tasks; Thinking, for complex work; and Pro, designed for maximum accuracy, best suited to challenging scenarios.

Fidji Simo, OpenAI's chief product officer, remarked that GPT-5.2 offers enhanced productivity in tasks such as coding, document analysis, and project planning.

According to TechCrunch, some employees suggested delaying the launch for further refinements, yet GPT-5.2 is aimed at improving OpenAI's enterprise opportunities.

The most notable aspect of GPT-5.2 is that it is designed to set new benchmarks in coding, math, science, and reasoning tasks, beating competitors like Gemini 3 in several areas.

OpenAI says that the Thinking variant makes 38% fewer errors than its predecessor, meaning it is more reliable for decision-making and research.

As OpenAI racks up its offerings, it seems to be focusing on reclaiming its lost ground in the AI space in the midst of increasing pressure from Google and changing market demands.