OpenAI has announced plans to remove several older artificial intelligence models from ChatGPT starting February 13, as the company continues its transition toward newer and more advanced large language models. The move will affect ChatGPT users worldwide and marks another milestone in OpenAI’s evolving product strategy.
The decision will affect ChatGPT users worldwide and includes the removal of GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini from the ChatGPT interface. While these models will no longer be selectable by regular users, OpenAI has confirmed that they will remain accessible through its application programming interface (API) for developers.
Among the models being retired, the most notable is GPT-4o, a system that gained a loyal following for its warm, conversational tone and human-like responses. Despite its popularity among a subset of users, OpenAI says GPT-4o now accounts for only 0.1 per cent of daily ChatGPT interactions, prompting the company to phase it out from the platform.
GPT-4o’s Removal Rekindles User Debate
The decision to remove GPT-4o has reignited debate among users, many of whom had expressed emotional attachment to the model’s personality and interaction style.
GPT-4o was briefly removed from the ChatGPT model picker in August, following the rollout of GPT-5. That earlier move triggered widespread backlash on social media, with users arguing that GPT-4o felt more empathetic, natural, and engaging than newer alternatives.
In response at the time, OpenAI reinstated GPT-4o and introduced customisable personality settings, allowing users to choose from seven predefined personalities or define their own through custom prompts. According to the company, this flexibility reduced the need to maintain multiple legacy models solely for stylistic preferences.
OpenAI Chief Executive Officer Sam Altman had previously stated that the company would provide advance notice before permanently retiring GPT-4o, after the earlier backlash. This time, OpenAI says it has taken steps to ensure users are informed ahead of the February deadline.
Which Models Are Being Retired?
In addition to GPT-4o, OpenAI confirmed it will also remove the following models from ChatGPT:
- GPT-4o
- GPT-4.1
- GPT-4.1 mini
- OpenAI o4-mini
These models will no longer be available for direct use within ChatGPT’s interface once the changes take effect.
However, OpenAI emphasised that this change applies only to ChatGPT’s consumer-facing platform. The models will remain accessible through OpenAI’s API, meaning developers can continue using them in apps, tools, and services without disruption.
At present, OpenAI says no changes are planned for the API, reassuring enterprise customers and software developers who rely on these systems.
Why OpenAI Is Phasing Out Older Models
The removal of older models reflects OpenAI’s broader effort to streamline ChatGPT and focus on systems that offer improved reasoning, efficiency, and scalability.
As newer models like GPT-5 gain traction, maintaining multiple legacy models becomes increasingly resource-intensive. According to OpenAI, most users have already migrated to newer options, making continued support for older models less practical.
By consolidating its offerings, OpenAI aims to:
- Improve platform performance
- Reduce operational complexity
- Focus research and safety efforts on fewer, more capable models
- Deliver more consistent user experiences
The company has repeatedly stated that innovation in AI requires regular transitions — even when older systems remain popular with a niche audience.
User Reactions: Progress vs Personality
While many users welcome faster, more capable models, others argue that technical improvements sometimes come at the cost of conversational warmth.
GPT-4o, in particular, became known for its friendly tone and emotional sensitivity, qualities some users felt were less pronounced in newer systems. Online forums and social media platforms have been filled with mixed reactions, ranging from acceptance of progress to disappointment over losing a “familiar” AI presence.
OpenAI’s introduction of personality controls appears to be an attempt to bridge that gap, allowing users to shape interactions without relying on older models.
What This Means for ChatGPT Users
For most users, the change is expected to have minimal impact, as usage data suggests the vast majority already rely on newer models. However, those who specifically preferred GPT-4o or GPT-4.1 will need to adjust their settings or experiment with personality customization options to replicate their preferred experience.
OpenAI has not indicated whether additional legacy models will be retired later this year, but the company has made it clear that model turnover will remain a regular part of ChatGPT’s evolution.
- Personality presets
- Custom instruction fields
- Prompt-based tone adjustments
The company believes these tools can replicate — or even improve upon — the experience users valued in older models.
Developers Largely Unaffected
While consumers adjust to the changes, developers are unlikely to feel any immediate impact.
OpenAI confirmed that all four models will remain fully supported through the API, allowing continued use in:
- Mobile and web applications
- Enterprise software
- Research projects
- Automation tools
This distinction between consumer-facing products and developer tools reflects OpenAI’s dual strategy of rapid innovation for users while maintaining stability for businesses.
A Sign of AI’s Rapid Evolution
The retirement of GPT-4o and other models serves as a reminder of how quickly artificial intelligence technology is advancing. In traditional software cycles, products can remain relevant for years. In AI, relevance can disappear in months.
As OpenAI continues to push toward more capable systems, users can expect:
- Faster model turnover
- More frequent updates
- Continued experimentation with personality and interaction styles
The company has not ruled out further retirements later in 2026.
Read More: Google defeats bid for billions of dollars of new penalties in US privacy class action
FAQs
When will GPT-4o and other models be removed from ChatGPT?
The models will be retired from ChatGPT on February 13, 2026.
Why is OpenAI removing GPT-4o if users liked it?
OpenAI says GPT-4o now accounts for only 0.1% of daily interactions, and maintaining older models slows progress toward more advanced systems.
Will developers lose access to these models?
No. GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini will remain available through OpenAI’s API.
Can users recreate GPT-4o’s personality?
Yes. ChatGPT now offers multiple predefined personalities and custom personality settings designed to reduce reliance on older models.
Are more model removals expected in the future?
OpenAI has not announced further retirements but has indicated that ongoing model updates and removals are part of its long-term strategy.
Conclusion
OpenAI’s decision to remove GPT-4o and three other legacy models from ChatGPT underscores the rapid pace of change in artificial intelligence development. While the move has sparked renewed debate among users attached to familiar systems, the company maintains that consolidation is essential to delivering better performance, safety, and innovation.
