OPENAI
GPT-5.2 takes over the hard parts of your job
OpenAI has launched GPT-5.2, its strongest model yet for professional work.
It improves on spreadsheets, presentations, coding, long documents, tool use and image understanding.
On GDPval, OpenAI’s performance measure covering 44 types of knowledge work, GPT-5.2 matches or beats industry professionals in 70.9% of tasks.
Companies like Notion, Box, Shopify and Databricks report clearer reasoning, stronger tool-calling and better performance in data, coding and document-heavy workflows.
GPT-5.2 Instant, Thinking and Pro start rolling out today for paid ChatGPT plans, with API access already live.
Per OpenAI, error rates are lower, long-context handling is stronger, and the model performs better on charts, interfaces and scientific diagrams.
Three key points:
GPT-5.2 reaches expert-level performance on many knowledge-work tasks.
Major gains in reasoning, coding, vision and long-context accuracy.
Available now across paid ChatGPT plans and the API.
Reads faster than you emotionally recover
Coding accuracy has also risen, with more reliable debugging and feature work.
Safety responses have been updated, and GPT-5.1 will remain available for three months.
API pricing is higher than before but still below other frontier models.
Long-context is so good it probably remembers things I’ve forgotten about my own life. - MG