Last Thursday, OpenAI announced ChatGPT-5.2, which offers better performance across the board. It’s also the company’s strongest model yet for science and math. In the announcement press release, OpenAI says that people using the latest model for work-related tasks will see the most benefit, not necessarily people who use ChatGPT day-to-day. 

«We designed GPT‑5.2 to unlock even more economic value for people,» the statement said. «It’s better at creating spreadsheets, building presentations, writing code, perceiving images, understanding long contexts, using tools, and handling complex, multi-step projects.»

Earlier this month, OpenAI CEO Sam Altman sent an internal «code red» memo to alert employees about the growing threat from competitors — specifically Google and its advanced Gemini 3 chatbot, released in November. The move echoes Google’s own «code red» memo from a few years ago, sent when ChatGPT launched and captured global attention. In just a few short years, the roles have reversed: Each company now sees the other as its main rival in the AI race.

Open AI also launched GPT-Image Model 1.5, an updated AI image model, on Tuesday to challenge Google’s Nano Banana Pro, which currently sits at the top of the list of the best AI image generators at CNET.

(Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.

There are three models of GPT-5.2. GPT‑5.2 Instant is designed to be fast and helpful for everyday activities. GPT‑5.2 Thinking is the most advanced version of GPT‑5.2 for professional, real-world tasks. And OpenAI says GPT-5.2 Pro is «our smartest and most trustworthy model yet, for difficult questions where a higher-quality answer is worth the wait.» OpenAI says that the entire ChatGPT-5.2 family provides meaningful upgrades from past versions for work and learning. 

For developers, ChatGPT-5.2 is designed to be a robust model for building agents, thanks to improvements in general intelligence, long-context understanding, agentic tool-calling and vision. (In AI and software development, «agents» are AI systems that can perceive information, such as user inputs, reason about what to do, and then take action, such as run code or operate software.)

The latest model of ChatGPT begins its rollout on Thursday, starting with paid plans, and is now available to all developers.