OpenAI has unveiled GPT-5.2, its most advanced AI model yet, just days after internally declaring a ‘code red’ over Google’s AI lead. Released on December 12, 2025, the model delivers significant improvements in professional reasoning, coding, long-document analysis and tool use, targeting enterprise and knowledge-work users.

BulletsIn

•  GPT-5.2 designed as OpenAI’s most capable model for enterprise-grade knowledge work and long-running agents.
•  Offers major boosts in coding, reasoning, spreadsheet creation, presentations, and image understanding.
•  Businesses using ChatGPT Enterprise earlier reported saving 40–60 min/day; GPT-5.2 expected to increase gains.
•  Comes in three versions: Instant (fast outputs), Thinking (deep reasoning), Pro (highest technical quality).
•  API rollout includes gpt-5.2, gpt-5.2-chat-latest, gpt-5.2-pro; pricing higher than 5.1 but more efficient overall.
•  On GDPval benchmark across 44 professions, GPT-5.2 Thinking matched/outperformed professionals in 70.9% tasks.
•  Sets new record on SWE-Bench Pro; better at debugging, feature-building, reviewing code and full engineering tasks.
•  Near-perfect long-context accuracy; can track info across hundreds of thousands of tokens with fewer hallucinations.
•  Stronger multi-step tool use; achieved 98.7% accuracy on Tau2 telecom-support scenarios.
•  Improvements also seen in scientific reasoning, graduate-level STEM tasks, mathematical proofs and visual analysis.