OpenAI Rolls Out GPT-5.2 With Enhanced Capabilities For Technical, Mathematical, And Scientific Applications
In Brief
OpenAI released GPT-5.2, its most advanced model to date, delivering expert-level performance across professional, technical, and scientific tasks with tiered versions for broader user and developer access.
Artificial intelligence research organization OpenAI reported the release of GPT-5.2, described as its most advanced model series to date for professional knowledge-focused tasks.
The model was developed to generate greater economic value by improving performance in areas such as spreadsheet creation, presentation development, coding, image interpretation, extended-context comprehension, tool use, and complex project execution.
GPT-5.2 is presented as setting new performance standards across a range of benchmarks, including GDPval, where it surpasses industry specialists on well-defined knowledge work spanning 44 occupations.
In this evaluation, GPT-5.2 Thinking reaches expert-level performance, matching or exceeding top professionals in the majority of assessed tasks. Human judges reported that the model outperformed or tied experts in more than two-thirds of GDPval comparisons, producing outputs such as presentations and spreadsheets with higher speed and at a fraction of the historical cost associated with human work.
Following testing, several companies, including Notion, Box, Shopify, Harvey, and Zoom, reported that GPT-5.2 shows advanced long-horizon reasoning and strong performance in tool-based workflows. Furthermore, Databricks, Hex, Triple Whale observed that the model performs particularly well in data-focused tasks, including agentic data science and detailed document analysis. Cognition, Warp, Charlie Labs, JetBrains, and Augment Codethe pointed out that GPT-5.2 delivers high-level results in software development activities such as interactive coding, code review, and identifying defects.
The model also establishes a new high score on SWE-Bench Pro, a demanding benchmark that evaluates practical software engineering across multiple programming languages. Its performance suggests enhanced reliability in debugging production systems, implementing features, refactoring extensive codebases, and completing end-to-end fixes. Early testing indicates notable gains in front-end development, including complex or nonstandard interface work and 3D components.
In scientific and mathematical domains, GPT-5.2 Pro and GPT-5.2 Thinking are now presented as the strongest models from the organization. Their results on benchmarks such as FrontierMath indicate advances in general reasoning and abstraction, qualities valuable for scientific activities involving coding, data analysis, and experimental planning. On the GPQA Diamond benchmark, which evaluates graduate-level, search-resistant questions, GPT-5.2 Pro records a score above 93%, with GPT-5.2 Thinking following closely.
GPT-5.2 Launches Across ChatGPT Paid Plans, Introducing Instant, Thinking, And Pro Variants
GPT-5.2 began rolling out in ChatGPT across paid subscription tiers, with API access becoming available simultaneously for developers. The release includes three variations designed for different levels of complexity: Instant for fast, straightforward work, Thinking for tasks requiring multi-step reasoning, and Pro for demanding research and extended projects. API usage is priced at $1.75 per million input tokens and $14 per million output tokens.
OpenAI highlighted that it is introducing GPT-5.2 versions gradually to maintain service consistency. Paid users will continue to have access to GPT-5.1 for a transitional period of three months before it is retired.
Disclaimer
In line with the Trust Project guidelines, please note that the information provided on this page is not intended to be and should not be interpreted as legal, tax, investment, financial, or any other form of advice. It is important to only invest what you can afford to lose and to seek independent financial advice if you have any doubts. For further information, we suggest referring to the terms and conditions as well as the help and support pages provided by the issuer or advertiser. MetaversePost is committed to accurate, unbiased reporting, but market conditions are subject to change without notice.
About The Author
Alisa, a dedicated journalist at the MPost, specializes in crypto, AI, investments, and the expansive realm of Web3. With a keen eye for emerging trends and technologies, she delivers comprehensive coverage to inform and engage readers in the ever-evolving landscape of digital finance.
More articles
Alisa, a dedicated journalist at the MPost, specializes in crypto, AI, investments, and the expansive realm of Web3. With a keen eye for emerging trends and technologies, she delivers comprehensive coverage to inform and engage readers in the ever-evolving landscape of digital finance.
