OpenAI plans to retire GPT-4o and several older ChatGPT models on 13 February 2026. The company says this will make it easier for users to choose models and allow them to focus on newer systems that people now use. For now, API access stays the same, and OpenAI promises to give advance notice before any future API changes.  

What Is Changing In ChatGPT? 

In a post on 29 January 2026, OpenAI announced it will remove ChatGPT 4.0, GPT 4.1, GPT 4.1 Mini, and OpenAI O4 Mini from ChatGPT. This update comes with the earlier retirement of GPT-5 Instant and GPT-5 Thinking. OpenAI also shared that it briefly brought back GPT-4.0 during the GPT-5 launch after some Plus and Pro users said they liked its dialogue, style, and creativity. The company used the feedback to improve newer models.  

• Effective date: 13 February 2026.  

• Models exiting ChatGPT: GPT-4o, GPT-4.1, GPT-4.1 Mini, and OpenAI’s 4 Mini, plus the earlier announced GPT-5 Instant and Thinking.  

• There is no change to the API right now. OpenAI says it will let users know before making any changes to API access.  

• After the update, existing chats and custom GPTs will automatically use GPT-5.2, according to the help center.  

• OpenAI says that ChatGPT (voice) and ChatGPT (images) will not change with this update since they use different models.  

Why It Matters for Users and Startups? 

For most users, removing these models means losing access to newer options that OpenAI says are better at reasoning and offer greater control over tone and temperament. The company notes that most people already use GPT-5.2, and feedback on GPT-4.0’s style has helped shape features such as the friendly presets and controls for warmth and enthusiasm in GPT-5.1 and GPT-5.2.  

For founders and product teams, the main effect is on operations. If your workflows rely on GPT-4.0’s voice or creative style, you could test them with GPT-5.2 and adjust prompts to keep your brand’s voice and user experience. OpenAI also says it is working to reduce unnecessary refusals and overly cautious responses while still improving safety, including protections for users under 18 in most markets.  

How will this affect developers and enterprises? 

Developers using the API are not affected at this time. OpenAI says the models powering ChatGPT will remain available through the API, and it will give advanced notice before retiring any API models. For enterprise and education customers, administrators can allow limited access to older models. However, organizations should see this as temporary, keep track of which models they use, and plan to move to the recommended newer models.  

Context and Background 

OpenAI earlier replaced GPT-4 in ChatGPT with GPT-4o and later brought back limited access to GPT-4o for paid users after hearing their feedback. This latest retirement makes the move to newer models official, aiming for better performance and more customizable conversation. OpenAI says only a small number of users still choose GPT-4o each day, and narrowing the options lets them focus on features most people want.  

Timeline And Following Steps 

Starting 13th February, 2026, these models will not be available in ChatGPT if you rely on GPT-4o’s tone for behavior, testing, or updating your prompts with GPT-5.2. Check style settings and review customer-facing tools, such as support scripts or developer onboarding assistance. API access stays the same for now, but it’s wise to watch for deprecation notices, lock in model versions if possible, and set aside time for testing.  

Retiring popular models can be disruptive, but OpenAI says this is part of a bigger plan to make ChatGPT more creative, reliable, and flexible for users with early planning and expedient adjustments. Most teams should handle the change smoothly.

Source: OpenAI to retire GPT-4o and older ChatGPT models on 13 February 2026