OpenAI launched its latest frontier model, GPT-5.2, on Thursday amid increasing competition from Google. The company pitched it as its most advanced model yet, designed for developers and everyday professional use. GPT-5.2 is coming to ChatGPT paid users and developers via the API in three distinct flavors. The Instant model is speed-optimized for routine queries like information-seeking, writing, and translation. The Thinking model excels at complex structured work like coding, analyzing long documents, math, and planning. The Pro model is the top-end option aimed at delivering maximum accuracy and reliability for difficult problems.
Fidji Simo, OpenAI’s chief product officer, stated that the model was designed to unlock even more economic value for people. She explained that GPT-5.2 is better at creating spreadsheets, building presentations, writing code, perceiving images, understanding long context, using tools, and linking complex, multi-step projects.
This launch lands in the middle of a competitive arms race with Google’s Gemini 3, which has been topping leaderboards across most benchmarks, though Anthropic’s Claude Opus-4.5 still leads in coding. Earlier this month, it was reported that OpenAI CEO Sam Altman issued an internal “code red” memo to staff amid declining ChatGPT traffic and concerns about losing consumer market share to Google. The memo called for a shift in priorities, stalling on commitments like introducing ads to instead focus on creating a better ChatGPT experience.
GPT-5.2 represents OpenAI’s push to reclaim leadership, even as some employees reportedly asked for the model release to be delayed for further improvements. Despite indications that OpenAI would focus more on consumer use cases with personalization, this launch strengthens its enterprise opportunities. The company is specifically targeting developers and the tooling ecosystem, aiming to become the default foundation for building AI-powered applications. Recent data shows enterprise usage of OpenAI’s tools has surged dramatically over the past year.
This comes as Gemini 3 has become tightly integrated into Google’s product and cloud ecosystem for multimodal workflows. Google also recently launched managed servers that make its services easier for AI agents to use.
OpenAI claims GPT-5.2 sets new benchmark scores in coding, math, science, vision, long-context reasoning, and tool-use. The company says this could lead to more reliable agentic workflows, production-grade code, and complex systems that operate across large contexts. These capabilities put it in direct competition with Gemini 3’s Deep Think mode, which targets math, logic, and science reasoning. On OpenAI’s own benchmark chart, the GPT-5.2 Thinking model edges out both Gemini 3 and Claude Opus 4.5 in nearly every listed reasoning test.
Research lead Adain Clark noted that stronger math scores are a proxy for a model’s ability to follow multi-step logic, maintain number consistency, and avoid subtle compounding errors. He stated these properties matter for a wide range of workloads like financial modeling, forecasting, and data analysis.
Product lead Max Schwarzer said GPT-5.2 makes substantial improvements to code generation and debugging and can walk through complex math step-by-step. Coding startups reportedly see state-of-the-art agent coding performance and gains on complex workflows with the new model. Beyond coding, he said GPT-5.2 Thinking responses contain 38 percent fewer errors than its predecessor, making it more dependable for decision-making, research, and writing.
GPT-5.2 appears to be a consolidation of OpenAI’s last two upgrades rather than a reinvention. GPT-5 laid the groundwork for a unified system with a router between fast and deep thinking modes. GPT-5.1 then focused on making that system more conversational and better for agentic tasks. The latest model turns up the dial on all those advancements for a more reliable foundation for production use.
For OpenAI, the stakes are high. The company has made massive financial commitments for AI infrastructure buildouts, commitments made when it had a first-mover advantage. Now, with Google pushing ahead aggressively, that bet might be driving the internal ‘code red’ urgency.
OpenAI’s renewed focus on reasoning models is also a risky strategic move. The systems behind its Thinking modes are more expensive to run than standard chatbots because they use more computing power. By doubling down with GPT-5.2, OpenAI may be setting up a cycle where it spends more to win on leaderboards, then spends even more to keep those high-cost models running at scale. The company is already reportedly spending more on compute than before, with most of its inference costs now paid in cash rather than through cloud credits.
Notably absent from this launch is a new image generator. Altman’s code red memo reportedly identified image generation as a key priority, especially after Google’s image model had a viral moment following its release. Google has since launched an upgraded version with better text rendering and realistic output, integrated across its products. OpenAI reportedly plans to release another new model in January with better images, improved speed, and better personality, though the company did not confirm these plans.
OpenAI also said it is rolling out new safety measures around mental health use and age verification for teens, but did not spend much of the launch event discussing those changes.