Core Insights - OpenAI officially announced the launch of the new flagship model GPT-5.1 series, aimed at enhancing the intelligence and conversational experience of ChatGPT [1][3] - The upgrade comes just months after GPT-5, which faced criticism for its lack of "human touch" and creativity [1] Model Features - The GPT-5.1 series includes two core models: GPT-5.1 Instant and GPT-5.1 Thinking, along with the new GPT-5.1 Auto intelligent feature [3] - GPT-5.1 Instant introduces an "adaptive reasoning" mechanism, becoming the first lightweight model to incorporate chain-of-thought (CoT), allowing it to autonomously assess problem complexity [3] - GPT-5.1 Thinking focuses on professional reasoning, doubling the response speed for simple tasks while extending the thinking period for complex questions to provide detailed answers [3] User Experience Enhancements - GPT-5.1 Auto automatically selects the most suitable model for each query, minimizing the need for manual selection [4] - The personalization features have been significantly upgraded, introducing eight dialogue style options and allowing advanced users to customize response simplicity, warmth, readability, and emoji usage frequency [4] - OpenAI's CEO and scientific VP highlighted the improvements in instruction adherence and adaptive thinking, stating that GPT-5.1 redefines human-computer interaction standards [4] Market Adoption - OpenAI reported that over 1 million businesses are directly using its services, with ChatGPT's weekly active users exceeding 800 million [5] - The number of commercial user seats for ChatGPT for Work has surpassed 7 million, reflecting a 40% increase in just two months [5] - OpenAI has introduced various tools for office environments to facilitate enterprise adoption, including integrations with platforms like Slack and Google Drive [5] Usage Growth - The Codex model for code generation and workflow automation has seen a tenfold increase in usage since August [6] - Notable users like Cisco and Carlyle have reported significant efficiency improvements through the use of OpenAI's models [6]