OpenAI Unveils GPT-5.2: A Major Upgrade Triggered by Rising Competition

OpenAI Launches GPT-5.2 Amid Rising Competition and Internal “Code Red” Push

OpenAI has unveiled GPT-5.2, its most advanced AI model to date, boasting major improvements across writing, coding, and reasoning capabilities. The release comes shortly after CEO Sam Altman internally declared a “code red,” directing the company to intensify its efforts to enhance ChatGPT in response to fierce competition.

Fidji Simo, OpenAI’s CEO of applications, clarified during a press briefing that the code red wasn’t about rushing GPT-5.2’s launch. Instead, it was introduced to refocus priorities and direct more resources toward strengthening ChatGPT. She noted that these additional efforts have been “helpful” in preparing the new model.

Once unmatched in the AI space following ChatGPT’s 2022 debut, OpenAI now faces serious challengers—especially Google, whose Gemini 3 model has gained significant traction with over 650 million monthly users. This competitive landscape has prompted OpenAI to pause some ambitious projects, such as implementing ads in ChatGPT, in favour of refining its core technology.

GPT-5.2 arrives in three versions—Instant, Thinking, and Pro—each tailored for different tasks. According to OpenAI, GPT-5.2 Thinking achieves the highest scores yet on the GDPval benchmark, outperforming human professionals in more than 70% of real-world tasks and completing them 11 times faster. The company also reports a 38% reduction in hallucinations compared to GPT-5.1.

Despite strong benchmark results, user experience remains a crucial challenge. Earlier model releases, including GPT-5, faced backlash for responses that felt colder or less engaging. Striking a balance between warmth and accuracy—without falling into excessive agreeableness—remains an ongoing focus.

OpenAI is also strengthening safeguards around mental health–related prompts. With over a million users reportedly discussing suicidal thoughts with ChatGPT each week, the company continues to refine how the model responds to sensitive situations. Additionally, OpenAI has begun deploying an age-prediction system to automatically apply content protections for users under 18, with “adult mode” expected to launch in early 2026.

As competition intensifies and user expectations rise, GPT-5.2 marks a pivotal step in OpenAI’s mission to enhance ChatGPT’s performance, safety, and overall experience.