JUST IN: OpenAI officially releases their first open source LLM, GPT-OSS. https://t.co/rL2S58OlbY ...

Amazon plans to make OpenAI’s new open artificial intelligence models available to customers https://t.co/r7serS4Klc ...

JUST IN: OpenAI officially releases their first open source AI model, GPT-OSS. https://t.co/NLblF9QRgX ...

Open AAI has delivered on their promise to release a state-of-the-art open-source model. This is GPTOSS. Now, I think the mystery model Horizon Alpha that was on Open Router is actually this open source model from OpenAI, although they have not confirmed that to me, but we do have an incredible model.Let me tell you about all the details. So, first it comes in two sizes, a 120 billion parameter version and a 20 billion parameter version. These are state-of-the-art openweight language models.Open weight. So ...

OpenAI is embracing open source for the first time in five years, with two new models you can download right now making their debut.Get the full rundown here: https://t.co/o1GGstdU8H https://t.co/VSun5P8gf5 ...

Model Overview - GPT-OSS is a state-of-the-art open-weights reasoning model with strong real-world performance comparable to O4-Mini [1] - The model can be run locally on personal computers or phones [1] - The company believes this is the best and most usable open model currently available [1] Impact and Availability - The model, resulting from billions of dollars of research, is being made available to the world to democratize AI access [2] - The company anticipates more good than bad will result from its release, citing performance on challenging health issues comparable to O3 [2] - The company has worked to mitigate serious safety issues, especially around biosecurity, with GPT-OSS models performing comparably to frontier models on internal safety benchmarks [2] - The company hopes this release will enable new kinds of research and the creation of new kinds of products, expecting a meaningful uptick in the rate of innovation [4] Strategic Vision - The company believes in individual empowerment, providing users with direct control and modification capabilities over their own AI, highlighting privacy benefits [3] - The company's mission is to ensure AGI benefits all of humanity [5] - The company is excited for the world to build on an open AI stack created in the United States, based on democratic values, available for free to all and for wide benefit [5]

OpenAI is releasing a pair of open and freely available AI models months after China's DeepSeek found success with a similar approach https://t.co/6AzFhFInXK ...

OpenAI launches two 'open' AI reasoning models | TechCrunch https://t.co/sV4gy22dV4 ...

另一方面，尽管OpenAI尚未官宣GPT-5的发布日期，但最近一系列不寻常的信号，已经把人们的胃口吊 了起来。 7月19日，OpenAI创始人阿尔特（300825）曼在X平台上发文："我们即将发布GPT-5。"没过几天，7月 24日，他在一档播客节目中首次提及GPT-5的内部测试体验，称其"令人震惊"，并表示"我们很快就会发 布它"。 当地时间周一，OpenAI宣布，其旗下ChatGPT本周的周活跃用户数将达到7亿，较3月的5亿有所增长， 这标志着其用户量同比激增逾4倍。该公司表示，这一数据涵盖了ChatGPT的所有人工智能产品，包括 免费版、PlusPro版、企业版、团队版和教育版，同时每日用户消息量已突破30亿条。与去年同期2.5倍 的同比增长率相比，当前增长速度还在加快。 第二，OpenAI将拥有更强的推理与解决问题的能力。中信证券研究8月1日发布的报告预计，较GPT-4单 次激活2770亿的规模，新一代模型GPT-5单次推理的算力消耗不会大幅增加，API价格或较GPT-4o仅小 幅增长甚至不增长，推理性价比提升一倍以上，这将带动AI应用盈利能力大幅改善。 第三，GPT-5将具备自主规划和执行任务的能力 ...

GPT-5被曝引入名为"通用验证器"的关键技术，或成OpenAI拉开竞争差距的"秘密武器"。其核心是"证明者-验证者游戏"的对抗性训练方法，通过模型间的博弈来自动化验证答案质 量，旨在攻克数学、创意写作等主观复杂领域的评估难题。该技术被指是前联合创始人Ilya"超级对齐"团队的"遗产"。 硬·AI 作者 | 龙 玥 编辑 | 硬 AI 在下一代大模型GPT-5备受期待之际，一项名为"通用验证器"的新技术正浮出水面，揭示了OpenAI可能用于拉开竞争差距的"秘密武器"。 OpenAI的"通用验证器"或将直接影响GPT-5模型的市场竞争力，8月4日据科技媒体The Information援引知情人士消息报道，这项技术已被应用于GPT-5的开发过 程中。 该技术的核心机制，被比作一场"证明者-验证者游戏"。 简而言之，它让一个AI模型扮演"验证者"的角色，去检查和评判另一个"证明者"模型生成的答案。通过这 种内部对抗和反馈，系统性地提升模型的输出质量。这一自动化流程旨在解决强化学习（RL）在创意写作等主观领域或数学证明等复杂领域难以验证的瓶颈。 OpenAI内部研究人员已在社交平台X上间接证实了相关方法的有效性。研 ...