OpenAI Releases GPT-5.2: A Game-Changer For Math, Coding, And Productivity

OpenAI announced the release of GPT-5.2 on Dec. 12, just a few days after CEO Sam Altman sounded a ‘Code Red’ alert amid stiff competition from Google’s Gemini 3. Altman shared on X, “It is a very smart model, and we have come a long way since GPT-5.1.”

In its official release note, OpenAI described GPT-5.2 as “the most capable model series yet for professional knowledge work.”

According to OpenAI, GPT-5.2 brings about more value for its users. Regular users of ChatGPT Enterprise save an average of 40 to 60 minutes daily, while heavy users save more than 10 hours per week. GPT-5.2 enhances spreadsheet generation, making presentations, coding, image understanding, understanding large contexts, usage of tools, and executing projects with multiple steps. It enhances these functionalities based on previous capabilities.

GPT-5.2 shows progress over GPT-5.1 on various benchmarks. In knowledge work, it attained 70.9% on GDPval compared with GPT-5 at 38.8%. In software engineering, it obtained 55.6% on SWE-Bench Pro (public) and 80.0% on SWE-Bench Verified, with better performance than GPT-5 at 50.8% and 76.3%, respectively. On science queries in GPQA Diamond, GPT-5.2 obtained 92.4%, an increase from 88.1%. On CharXiv Reasoning with Python, it attained 88.7%, an upgrade from 80.3%. On competitive mathematics problems, GPT-5.2 obtained 100.0% on AIME 2025 and 40.3% on FrontierMath Tier 1-3 compared with 94.0% and 31.0%, respectively, for GPT-5.

A number of businesses have commended the capabilities available in GPT-5.2. According to OpenAI, Notion, Box, Shopify, Harvey and Zoom were impressed with its reasoning on a long horizon and tool-calling abilities. Databricks, Hex and Triple Whale appreciated the AI tool’s superiority in data science and document processing. Cognition, Warp, Charlie Labs, JetBrains and Augment Code were happy with its improvements within interactive coding, reviews and debugging.

GPT-5.2 will be available on the ChatGPT Instant, Thinking, and Pro plans beginning today for paid accounts, as well as available today via the API for all developers. On the whole, GPT-5.2 shows remarkable progress in general intelligence, large-context understanding, utilisation of tools and vision capabilities, making it possible to carry out complex tasks effectively compared to all previous models.