OpenAI launches GPT 5.2 to counter Gemini 3, after CEO Sam Altman declared 'code red'

OpenAI has officially launched its latest large language model, GPT-5.2, positioning it as the company's most advanced offering yet for professional knowledge work and advanced reasoning. The release on 11th December comes hot on the heels of intensified rivalry with Google, whose Gemini 3 model had recently prompted OpenAI CEO Sam Altman to issue an internal "code red."

GPT-5.2 has been rolled out across three distinct tiers: Instant, Thinking, and Pro, designed to cater to a broad spectrum of computational needs, from rapid information retrieval to highly complex, multi-step projects. The company claims significant performance enhancements in core areas critical for enterprise use.

Notably, the model demonstrates cutting-edge mathematical capability. OpenAI reported that the mid-range Thinking version achieved a record score on the challenging FrontierMath benchmark, correctly solving 40.3% of college-level problems. In a significant case study, the top-tier GPT-5.2 Pro was credited with autonomously solving a simple version of an open problem in statistical learning theory, with human researchers focused solely on verification.

Beyond pure intellect, the model features vastly improved performance in software engineering, registering a new state-of-the-art score on the rigorous SWE-Bench Pro coding evaluation. It is also significantly better at understanding charts and figures within scientific papers, improving accuracy on the CharXiv Reasoning benchmark by over 8% compared to its predecessor.

OpenAI is marketing GPT-5.2 as a reliable work partner, with the Thinking version reportedly achieving a win rate of over 70% when judged against top industry professionals on well-specified knowledge work tasks such as spreadsheet creation and presentation building. The model is now available to ChatGPT subscribers and via the developer Application Programming Interface (API), marking a determined effort by the firm to reclaim momentum in the rapidly evolving artificial intelligence landscape.