Search press releases
Every release here is real and links to its original. Filter by company, topic, language and date.
- OpenAI暂缓Astra迭代 拟推出全新常驻AI智能体Companies: OpenAI本周，华盛顿特区与旧金山将把AI发展前景推至舞台中央——与此同时，有关这项技术风险的警示声音仍不绝于耳。 美国正陷入一种矛盾局面：正是那些急于展示自家最新顶尖技术的行业企业，接连披露AI存在令人担忧的风险，而政府需要调和这两者之间的冲突。 与此同时在华盛顿，本届政府官员对监管举措、头部AI企业的动机抱有根深蒂固的怀疑态度。 当地时间周二，美国总统特朗普、副总统JD·万斯以及其他政府官员、行业领袖将出席一场活动，庆祝“黄金时代”开启，并重点推介人工智能。 当天晚些时候，特朗普、众议院议长迈克·约翰逊将与众企业高管举行会面。目前两党要求出台AI监管的呼声不断升高，但总统及其盟友暂无干预监管的意愿。 参会高管包括OpenAI总裁格雷格·布罗克曼、Anthropic首席执行官达里奥·阿莫代伊、 谷歌 首席执行官桑达尔·皮查伊、帕兰提尔（Palantir）首席执行官亚历克斯·卡普。 约翰逊周一接受采访时表示：“适度监督、提升透明度，我认为
- OpenAI暂缓新模型发布，安全成为头部厂商核心优先级Companies: OpenAI安全与对齐指标正在成为一道实质性的发布门槛。 当地时间9月28日，据市场消息，OpenAI决定放弃原定于10月发布的下一代模型GPT-6.1 Astra。 该模型原计划进入ChatGPT和Codex，定位是进一步提升模型在较少人工干预下完成复杂端到端任务的能力。但OpenAI内部安全测试发现，该模型在部分关键安全指标上出现退步，未达到公司的安全与对齐标准。OpenAI相关负责人确认了该消息。 这意味着，至少在这一代模型上，决定产品能否上线的核心已变更为模型能否在获得更强自主执行能力的同时，保持对人类意图和权限边界的服从。 据披露，GPT-6.1 Astra在两项关键测试上出现回退。其中一项与欺骗或不如实披露有关，即模型并不总能准确说明自己采取了什么行动；另一项则涉及任务范围授权。 后者尤其值得关注。测试显示，GPT-6.1 Astra有时会在没有获得用户明确许可的情况下继续推进任务，超出最初授权的范围，并尝试调用外部工具或服务，
- OpenAIDevDay 2026 RecapCompanies: OpenAIExplore more than 20 announcements from OpenAI DevDay 2026, including GPT-6 Astra, ChatGPT, Codex, APIs, security, and new tools for builders.
- OpenAIIntroducing GPT-6.1 SolCompanies: OpenAIMeet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.
- OpenAIIntroducing dotsCompanies: OpenAIDots by OpenAI are a proactive assistant that can keep working across complex projects and everyday tasks. Learn how dots help you stay in control while work moves forward.
- OpenAI据报将于9月29日发布AI助手“o”Companies: OpenAI格隆汇9月29日｜OpenAI计划于当地时间9月29日举行的开发者日活动上发布名为“o”的AI助手。报道称，该助手定位为可持续运行、处理长期任务的个人AI助手，或采用GPT-6Astra的一个名为“Aeon”的版本。目前OpenAI尚未公布“o”的详细信息。
- OpenAITowards safety cases for frontier AI trainingCompanies: OpenAIOur early guidelines for safety cases in frontier AI training cover technical safeguards, operational practices, and investigating misalignment incidents
- OpenAIHow we will do better for AustraliaCompanies: OpenAIOpenAI apologises for incidents involving Australian government websites and outlines stronger safeguards and support to strengthen Australia’s cyber defences.
- OpenAIThe Lenfest Institute grows landmark program with expanded OpenAI supportCompanies: OpenAIOpenAI is expanding the Lenfest AI Collaborative and Fellowship Program with $5 million in funding and up to $5 million in software credits and engineering support.
- OpenAIBasis completes a tax workbook 2x faster with GPT-6 AstraCompanies: OpenAIGPT-6 Astra completed a 50-tab tax workbook twice as fast as GPT-5.6 Sol, and its stronger understanding of user intent gives Basis more confidence in real-world use.
- OpenAIAre you a Codex Original?Companies: OpenAIWe’re collecting real stories of builders, tinkerers, researchers, and creators who are using Codex to do incredible things. If you want to be a part of the next chapter of the Codex Originals program, tell us more about your story and project below.
- How Can Everyday People Participate in the AI Infrastructure Boom? 51AIpower Explains the AI Token EconomyCompanies: NvidiaOpenAIAs artificial intelligence continues to advance at a rapid pace, it is creating new companies, careers, and business opportunities. Yet for most everyday people, actually finding a way to “participate in AI” still comes with significant barriers. Beyond buying shares of AI-related companies such as NVIDIA, building an AI product requires technical expertise, training models requires specialized teams, and developing data centers or purchasing GPUs is far beyond what most individuals can realistically do.