GPT-5.4 Mini Rollout

OpenAI is rolling out GPT-5.4 mini to ChatGPT across all user tiers. Free and Go users can access GPT-5.4 mini directly via the "Thinking" feature in the plus menu. For Plus, Pro, and other paid users, GPT-5.4 mini serves as a rate-limit fallback for GPT-5.4 Thinking, ensuring continued access to reasoning capabilities during high-usage periods. Enterprise customers retain the option to default Auto routing to GPT-5.4 mini if preferred.

Note that GPT-5.4 mini will not appear as a selectable model in the model picker, and GPT-5 Thinking mini will be retired in 30 days, streamlining the available model options.

GPT-5.4 Thinking Improvements

GPT-5.4 Thinking incorporates significant enhancements across multiple dimensions:

• Mid-response adjustments: Users can now see an upfront plan of the model's thinking and adjust course while it's working, enabling faster convergence to the desired output without additional turns.
• Deep web research: Particularly improved for highly specific queries, giving users more accurate and targeted research results.
• Context management: Better maintains context for longer thinking sessions with improved context window management, enabling the model to think for longer without losing track of the conversation.
• Speed and quality: Together, these improvements deliver higher-quality answers that arrive faster and remain more relevant to the task at hand.

Additional Updates

OpenAI has also deprecated legacy models (GPT-5.1, GPT-4o, GPT-4.1, and o4-mini) as previously announced, with existing conversations automatically migrating to current equivalents. Minor updates to GPT-5.3 Instant include improved tone, reduced teaser phrasing, and better conversational flow.