The news: OpenAI’s GPT-5.2 is its answer to advancing models from Gemini and what the company calls its best model yet for professional use.

The latest model improves spreadsheet creation, presentation building, code generation, understanding of images and long contexts, tool use, and handling of complex projects, per an OpenAI blog post.

The update aims to remix and improve on ChatGPT’s two latest upgrades, GPT-5 and GPT-5.1. As with those, GPT-5.2 can toggle between multiple modes depending on the task and users’ willingness to wait for a response:

• Instant. This quick-thinking version is optimized for searching, writing, and translating.
• Thinking. It works well for complicated projects like coding, planning, or math.
• Pro. The “smartest and most trustworthy” tier, per OpenAI, is intended for deeply complex or high-stakes queries when accuracy is more important than speed.

What sets it apart: GPT-5.2 is designed to be more reliable for production and development work, addressing earlier models’ inconsistencies with heavier tasks.

• It shows better adherence to complex instructions and more willingness to attempt hard tasks.
• It’s also more consistent in tool use, making it a stronger fit for enterprise automation and collaborative workflows where precision and repeatability are important.
• While it has greater long-context thinking, Pro has the tendency to “think forever and still fail,” per a review from OthersideAI CEO Matt Shumer.

Zooming out: The model follows CEO Sam Altman’s “code red” declaration to prioritize ChatGPT advancements amid rising competition from Google’s Gemini 3, Anthropic’s Claude 4.5, and Perplexity. The company stated that GPT-5.2 had been in the works for months and denied moving up its release because of the code red, per Wired.

Between August and November, Gemini’s global website visits doubled while ChatGPT’s rose about 1%, per Sensor Tower data cited by The Information. Gemini’s monthly active users grew about 30% to 346 million while ChatGPT’s increased about 5% to 810 million.

Gemini may still compete with GPT-5.2 on raw speed and processing big jobs, but GPT-5.2 could pull ahead on reliable reasoning and how well it works with other tools, which could make it a stronger default for day-to-day enterprise work.

What brands should do: As brands expand past using AI for image generation and into higher-level tasks, such as financial modeling, market research, or campaign ideation, experiment with GPT-5.2 by:

• Testing each mode to see which offers the most dependable outputs for analytics, planning, and creative development.
• Upskill teams on tool-based workflows.
• Build AI-assisted production pipelines for tasks where accuracy and repeatability matter, such as forecasting and content deployment.

This content is part of EMARKETER’s subscription Briefings, where we pair daily updates with data and analysis from forecasts and research reports. Our Briefings prepare you to start your day informed, to provide critical insights in an important meeting, and to understand the context of what’s happening in your industry. Non-clients can click here to get a demo of our full platform and coverage.