GPT-5.1 Upgrade: How ChatGPT Gets Warmer and Smarter
▼ Summary
– OpenAI released GPT-5.1 with two new models, GPT-5.1 Instant and GPT-5.1 Thinking, to enhance ChatGPT’s user experience.
– GPT-5.1 Instant is designed for simpler prompts, offering warmer conversations, better instruction-following, and adaptive reasoning for accuracy.
– GPT-5.1 Thinking focuses on complex tasks by adjusting thinking time based on prompt complexity, improving response clarity and reducing jargon.
– Users can now personalize ChatGPT with new tone options like Professional, Candid, and Quirky, and adjust characteristics such as conciseness and warmth.
– The models are rolling out first to paid subscribers, with GPT-5.1 becoming the default for Enterprise and Edu plans, while legacy models remain available temporarily.
The latest upgrade to ChatGPT introduces GPT-5.1, a significant enhancement designed to deliver smarter and warmer interactions. This release addresses user feedback by restoring the ability to choose specific models and personalize the chatbot’s behavior, offering a more tailored experience. With two new models, GPT-5.1 Instant and GPT-5.1 Thinking, alongside expanded customization options, OpenAI aims to make conversations feel more natural and responsive to individual needs.
GPT-5.1 Instant serves as the go-to model for everyday queries, where quick, straightforward answers are usually sufficient. It has been refined to adopt a warmer and more conversational tone by default, making exchanges feel less robotic and more engaging. Early testers noted its playful yet practical responses, highlighting improvements in how well it follows instructions. For the first time, this model incorporates adaptive reasoning, allowing it to pause and consider more difficult prompts before replying. This results in greater accuracy on complex tasks like mathematical problems or coding challenges, while maintaining speed for simpler requests. Internal benchmarks show notable gains on standardized tests such as AIME 2025 and Codeforces.
For tasks demanding deeper analysis, GPT-5.1 Thinking steps in. Built to handle intricate problems, it dynamically adjusts how much “thinking” time it allocates based on the complexity of each question. This means it can respond nearly twice as fast for quick tasks and take twice as long for tougher ones, striking a balance between speed and thoroughness. The model also produces clearer, less jargon-filled answers, establishing it as OpenAI’s most advanced offering to date.
Both models are now being released in stages, with priority access for paid subscribers on Pro, Plus, Go, and Business plans. Free and logged-out users will gain access shortly after. GPT-5 Auto remains available for those who prefer automatic model selection, while the new options cater to users wanting specific performance or a particular conversational style. Enterprise and Education plans will see an early-access toggle before GPT-5.1 becomes the default. Legacy GPT-5 models will remain accessible for three months, letting subscribers compare performance between versions.
Personalization receives a major boost with this update. Users can now select from additional personality tones including Professional, Candid, and Quirky, alongside updated versions of Default, Friendly, and Efficient. Previously introduced options like Cynical and Nerdy remain available in settings. Beyond tone, new sliders allow fine-tuning of response characteristics such as conciseness, warmth, and readability. One popular addition addresses emoji usage, letting people control how often they appear in replies. An experimental feature will even adjust preferences automatically based on chat interactions, eliminating the need to manually update settings.
Custom instructions are now better understood by GPT-5.1, enabling more consistent personalization across all chats, including those already in progress. This marks a shift from earlier versions where changes only applied to new conversations. Together, these updates make ChatGPT not only more capable but also more adaptable, reflecting a clear focus on user-driven customization and improved conversational quality.
(Source: ZDNET)
