OpenAI Negotiates Licensing Content with CNN, Fox, and Time Magazine to Enhance ChatGPT Product


OpenAI publicly questioned the SWE-Bench Pro benchmark, pointing out that about 30% of its 731 test tasks have evaluation flaws. The benchmark, launched by Scale AI, is an industry authority for measuring large model programming capabilities. However, OpenAI warned that the passing rate of cutting-edge models has surged from 23.3% to 80.3% within 8 months, which is unusually fast, indicating doubts about the reliability of the evaluation.
Cloudflare and OpenAI launch a research pilot to leverage Cloudflare's global network data, enhancing real-time accuracy of AI search. This helps OpenAI models capture the latest web updates, addressing the key challenge of information access in generative AI.....
OpenAI launched a new voice interaction based on the GPT-Live architecture, transforming AI assistants from walkie-talkie-style communication to full-duplex human-like conversations. Compared to the advanced voice mode in 2024, the new model enables parallel processing of listening, speaking, and thinking, breaking free from the constraints of traditional voice technology.
OpenAI is recruiting investment banking experts to join its San Francisco AI team as 'subject matter experts' to 'teach' rather than use AI. The position offers an annual salary of up to $205,000, along with equity incentives. This move comes as OpenAI has secretly been preparing for an IPO. The initiative highlights OpenAI's active effort to attract financial expertise, aiming to deepen its presence in the investment sector.
Ex-OpenAI researcher Tian Yonglong joins Tencent to lead multimodal & vision-language model R&D at Hunyuan. Following Yao Shunyu, another ex-OpenAI talent joins Tencent Hunyuan, where they previously worked together. Yao joined in Sept 2025 as Chief AI Scientist, reporting to the president.....