OpenAI Announces GPT-5.2 Models With Major Productivity Upgrades, Aims To Challenge Gemini 3 Pro

OpenAI has announced the GPT-5.2 family of AI models, marking the second major upgrade to its fifth-generation series. The company says the new versions bring better performance for everyday tasks such as making spreadsheets, preparing presentations, and managing work schedules. OpenAI also shared benchmark results showing GPT-5.2 outperforms Google’s Gemini 3 Pro in several tests. The rollout will start with paid ChatGPT users before expanding to others. Also Read: Hackers Push Fake ChatGPT Atlas Browser Via Search Ads To Steal Passwords: Report

GPT-5.2 Model Availability

The GPT-5.2 lineup – Instant, Thinking, and Pro – will first be released to paid plans like ChatGPT Go, Plus, Pro, Business, and Enterprise. Free users and those on Edu accounts will get the update later. OpenAI also confirmed that GPT-5.1 will remain accessible under legacy models for the next three months, after which it will be discontinued. Also Read: Adobe Brings Photoshop, Express And Acrobat Support To ChatGPT For All Users

Although the main announcement did not directly mention Gemini 3 Pro, OpenAI CEO Sam Altman posted a set of benchmarks showing areas where GPT-5.2 has an edge over Google’s flagship model. Also Read: You Can Now Edit Images, PDFs And Designs Directly In ChatGPT: Here's How To Use

Key Improvements in GPT-5.2

One of the biggest upgrades comes in long-context handling. Earlier models struggled to retain accuracy when working with larger documents or long conversations. GPT-5.2 aims to handle longer inputs more effectively. This should help the model keep track of information while going through large reports, detailed documents, or longer back-and-forth chats without losing earlier details.

OpenAI says the model now performs better with multi-step instructions, including generating spreadsheets, writing or fixing code, preparing slide decks, and solving structured problems. The company adds that users should see fewer errors in zero-shot responses, reducing the need for repeated corrections.

GPT-5.2 also improves tool-calling capabilities, allowing it to interact more accurately with external apps or functions – such as triggering APIs, operating spreadsheet tools, or running code.

In terms of vision, the model can now interpret dashboards, diagrams, screenshots, and other technical visuals with better accuracy. OpenAI says ChatGPT responses will also feel more consistent and structured while keeping a natural conversational style.

Developers will be able to access GPT-5.2 through the API. Pricing starts at $1.75 per million input tokens and $14 per million output tokens, while the GPT-5.2 Pro tier is priced at $15 per million inputs and $168 per million outputs.