来自世界各地可信来源的关于 Intelligence 的最新新闻和动态。
In the middle of the war, Israel’s Channel 12 told the spokesperson of the Khatam al-Anbiya base that it is artificial intelligence and not a real human; now today, this guy didn’t take part in the ceremony, even though both Vaahidi and Abdollahian, the commanders of the IRGC, attended. It seems, really, it’s AI. .𓄂
A Swedish television network reported that one of the employees of Sweden’s immigration agency was dismissed from his job on suspicion of transferring confidential information to Islamic Republic intelligence agents. The “TV4” network reported that the employee had access to information related to opponents of the Islamic Republic, Iranian refugees, and in some cases Afghan refugees residing in Sweden, and is…
The Royal Swedish Academy of Sciences has awarded the 2025 Nobel Prize in Chemistry to David Baker, Demis Hassabis and John Jumper for advances in computational protein design and protein structure prediction. The prize recognizes work that has helped scientists better understand how proteins form and function, a field with major implications for medicine, biotechnology and drug discovery. Baker was honored for…
OpenAI has introduced o3-mini, a new reasoning-focused AI model designed to deliver strong performance while using less computational power. The company says the model is optimized for efficiency, aiming to reduce inference latency without sacrificing quality on demanding tasks.Built for Math, Science, and CodingAccording to OpenAI, o3-mini performs competitively on benchmarks in math, science, and coding, areas…
OpenAI has officially unveiled its o1 series of AI models, a new generation designed to spend more time “thinking” before answering. The company says the approach improves performance on complex tasks that require careful reasoning, including science, coding, and mathematics.Built for harder problemsUnlike earlier models that prioritize fast responses, o1 is tuned to work through multi-step challenges with greater…
OpenAI has introduced o1 Pro, a new reasoning model designed to deliver stronger performance on demanding tasks in science, coding, and analysis. According to the company, the model outperforms its predecessor, o1, on key benchmarks including AIME and GPQA, with a reported 94% score on AIME math problems. The release underscores OpenAI’s push to improve advanced reasoning capabilities as competition intensifies…
OpenAI has introduced o1-pro, a new reasoning model the company says delivers its best performance yet on complex problem-solving tasks. In benchmark testing, o1-pro outperformed the standard o1 model on challenging evaluations including AIME math, where it scored 93% versus 79%, and GPQA science, where it reached 84% compared with 74% for o1. The results position o1-pro as a more capable option for users who need…
OpenAI on December 5, 2024, introduced o1-pro, a new version of its reasoning model family that the company says delivers its strongest performance yet on complex analytical tasks. Built on the o1 series, the model is designed to improve results in areas such as mathematics, coding, and scientific problem-solving, where multi-step reasoning is critical. According to OpenAI, o1-pro has posted top scores on…
OpenAI has released o1-pro, a new reasoning model designed to handle complex tasks with greater accuracy and depth than its predecessor, o1. The company says the model performs especially well on demanding math and coding benchmarks, reflecting stronger step-by-step reasoning capabilities. According to OpenAI, o1-pro achieved a 93% score on AIME 2024 and 75% on GPQA Diamond, two widely watched tests used to measure…
OpenAI has launched o1-pro, a new reasoning model it says is its most capable yet, with availability beginning December 5, 2024 through ChatGPT Pro and the API. The company says the model is designed to handle complex tasks that require deeper step-by-step reasoning, and early results suggest strong performance on demanding benchmarks in math, coding, and advanced science problems. According to OpenAI, o1-pro…
OpenAI has released o1 Pro, a new reasoning-focused AI model designed to deliver stronger performance on complex tasks such as mathematics, coding, and multi-step problem solving. The company says the model outperforms earlier versions on demanding benchmarks and can reach what it describes as PhD-level capability in certain domains. The new model is available immediately to ChatGPT Pro users and through OpenAI’s…
OpenAI has introduced o1-pro, a new reasoning model the company says delivers its strongest performance yet on complex tasks such as mathematics and coding. The release marks another step in the rapidly intensifying competition among leading AI developers, as firms race to build models that can solve harder problems with greater reliability. According to OpenAI, o1-pro improves on earlier versions by handling…
