OpenAI officially introduced its new GPT-5.2 model this Thursday (11), the latest engine behind ChatGPT. The company describes it as its most capable assistant-focused language model so far, built to handle workplace tasks and long-running agent workflows with greater accuracy and efficiency.
Compared to GPT-5.1, released only a few weeks ago, the new model shows stronger performance in generating spreadsheets, creating slide decks, writing code, and hitting higher scores across key benchmarks.
Keeping an eye on Google
The timing of the release also fits into the broader competition between OpenAI and Google for leadership in the AI space. Google recently unveiled its Gemini 3 line, which quickly took top marks in several intelligence evaluations and gained extra attention thanks to its Nano Banana imaging model.
While Google has been gaining momentum, OpenAI has faced pushback on some of its latest updates. GPT-5 delivered a significant leap in capability but drew criticism for introducing an automated system that evaluates the complexity of each user request.
According to The Wall Street Journal, CEO Sam Altman even issued a “code red,” pushing the company to accelerate improvements to ChatGPT as rivals grow stronger. GPT-5.2 appears to be the first major response to that pressure.
What’s New in GPT-5.2?
OpenAI highlights several upgrades aimed at making the model feel more like a professional-grade assistant:
- Better support for building and analyzing spreadsheets, slide presentations, and code projects
- Improved understanding of visual content, including images, PDFs, dashboards, and data tables
- Stronger reasoning over long documents and extended context
- Fewer hallucinations and more reliable responses
GPT-5.2 comes in three versions: Instant (optimized for speed), Thinking (built for deep reasoning tasks), and Pro (the highest-performance option). For now, only paid ChatGPT subscribers can access the new model.
, the latest engine behind ChatGPT. The company describes it){kind=link}