OpenAI Launches GPT-5.1 With Enhanced Reasoning Features
OpenAI has announced the release of GPT-5.1, introducing two upgraded models that bring significant improvements in both intelligence and user experience. The company launched GPT-5.1 Instant and GPT-5.1 Thinking, building on the foundation of the GPT-5 series with enhanced capabilities and more natural communication styles.
GPT-5.1 Instant represents the most widely used model in the ChatGPT platform and now features a warmer, more conversational default tone. The model demonstrates improved instruction following, allowing it to respond more accurately to user questions. A notable addition is adaptive reasoning, which enables the model to pause and think through complex questions before responding, while maintaining quick response times for simpler queries. This advancement has led to measurable improvements in mathematical and coding evaluations, including AIME 2025 and Codeforces benchmarks.
The GPT-5.1 Thinking model has been optimized to adjust its processing time based on question complexity. The system now allocates more time to difficult problems while responding faster to straightforward requests. The model also produces clearer responses with reduced technical jargon and undefined terms, making it more accessible for professional and technical applications. Like its Instant counterpart, GPT-5.1 Thinking features a warmer and more empathetic communication style.
The rollout begins today for paid subscribers across Pro, Plus, Go, and Business plans, followed by free and logged-out users. Enterprise and Education plans receive a seven-day early access period with an optional toggle feature. After this window, GPT-5.1 will become the default model across the platform. OpenAI plans a gradual deployment over several days to maintain system performance and stability.
Both models will be integrated into the API later this week. GPT-5.1 Instant will be available as gpt-5.1-chat-latest, while GPT-5.1 Thinking will be released under the GPT-5.1 API designation. Both API versions will include adaptive reasoning capabilities.
The previous GPT-5 models will remain accessible through a legacy models dropdown for paid subscribers for three months, allowing users time to compare performance and adjust workflows. This transition period applies specifically to GPT-5 models and does not affect other legacy model availability. OpenAI has committed to providing clear communication and advance notice for future model sunset periods.
Alongside the model upgrades, OpenAI has expanded customization options for ChatGPT interactions. The company has refined its preset tone options to include Default, Friendly, Efficient, Professional, Candid, and Quirky settings. The original Cynical and Nerdy options remain available. These personality settings work across all models in the platform.
OpenAI is also testing granular control features that allow users to adjust specific characteristics such as response conciseness, warmth, scannability, and emoji frequency. The system can proactively suggest preference updates during conversations when it detects requests for particular tones or styles. All personalization settings now apply immediately across all conversations, including ongoing chats, representing a change from the previous system where settings only affected new conversations.
The company has labeled this release as GPT-5.1 to indicate meaningful improvements while remaining within the GPT-5 generation. Future incremental updates to GPT-5 will follow the same numerical designation pattern. A system card addendum detailing the safety approach for GPT-5.1 has been published alongside the release.
