OpenAI has officially retired GPT-4o as of February 13, 2026, completing the transition to GPT-5.2 as the default model across all ChatGPT subscription tiers — Free, Plus, and Pro.
The retirement marks the definitive end of the GPT-4 era, which began in March 2023 and served as the backbone of ChatGPT for nearly three years. GPT-4o, launched in May 2024 as a faster and more efficient multimodal variant, was the last model in the GPT-4 family still available to users.
GPT-5.2, which launched on December 10, 2025, brings substantial improvements across the board. The model features a 400K context window with 128K output tokens and comes in three variants: Instant (fast responses), Thinking (chain-of-thought reasoning), and Pro (maximum compute for complex tasks).
Key improvements over GPT-4o include:
3.2x better performance on coding benchmarks (HumanEval, SWE-bench)
Native tool-calling and agentic workflows built into the architecture
Significantly improved long-context understanding and retrieval
Better vision capabilities with higher accuracy on image analysis
State-of-the-art results on GDPval, outperforming industry professionals across 44 occupations
For API users, GPT-5.2 is priced at $1.75 per million input tokens and $14 per million output tokens, with a 90% discount on cached inputs. This represents roughly a 40% cost reduction compared to GPT-4o for most workloads due to efficiency improvements.
Free tier users now get access to GPT-5.2 Instant with approximately 10 messages every 5 hours. Plus subscribers ($20/month) unlock GPT-5.2 Thinking mode with 5x higher usage limits, while Pro users ($200/month) get unlimited access to GPT-5.2 Pro.
OpenAI CEO Sam Altman noted in a blog post that the transition was smoother than expected, with over 92% of enterprise customers already migrated before the cutoff date. The company says GPT-4o API endpoints will continue to function for existing integrations through March 31, 2026, giving developers additional time to update their systems.