Skip to content

Translation in progress Some entries still include text in the original language.

OpenAI · News

Oct 12 · Monday

OpenAI · News

Latest publications and developments
OpenAI

Research topics

Prompt injectionAgent safetyDeception and schemingReward hackingChain-of-thought monitoringInterpretabilityAI controlBackdoors and poisoning隐私与数据泄露Supply-chain securityDangerous capabilitiesAI 与网络攻防开源模型安全Red teaming

Companies and institutions

OpenAIAnthropicGoogle DeepMindGoogle ResearchGoogle 安全MetaxAIMicrosoftNVIDIADeepSeek阿里巴巴 · Qwen智谱 · GLM月之暗面 · Kimi字节跳动腾讯蚂蚁集团上海人工智能实验室UK AISIUS CAISIMETRApollo ResearchRedwood ResearchFAR.AIPalisade Research

Reports and frameworks

System CardFrontier safety frameworks
TodayOct 12Mon
  1. The DecoderThe Decoder · 63

    DeepMind AlphaProof Nexus solves dozens of open mathematical problems

    Google DeepMind's AlphaProof Nexus solves dozens of open mathematical problems, and the related research was published in Science and passed peer review. The system solved 9 of 353 open Erdős problems, two of which had been open for 56 years; it also proved 44 conjectures in the OEIS integer sequence database, solved an open problem in algebraic geometry, and improved a known bound in min-max optimization used to train some AI models. The system is based on Gemini 3.1 Pro and uses the Lean formal method.

Oct 11Sun
  1. The DecoderThe Decoder · 46

    Study: Multi-agent teams are costly but offer limited quality gains

    Evaluation company Vals AI compared single-agent and multi-agent teams of GPT-6 Sol and Claude Opus 5.5 on Vibe Code Bench, finding that team costs were 1.8 to 5.1 times those of a single agent, with a statistically significant improvement in only one of four comparisons. That improvement occurred at GPT-6 Sol's medium reasoning intensity, where the team scored 7.3 points higher; at the highest reasoning intensity, the team setup offered no practical advantage for either model. Anthropic's own tests with Opus 5.5 also showed limited gains from adding more agents.

  2. 量子位量子位 · 25

    Silicon Valley AI computing power competition intensifies, GPUs become hard currency

    The competition for AI compute in Silicon Valley continues to heat up. According to QbitAI, Anthropic has partnered with SpaceX, gaining over 300 megawatts of new capacity and 220,000 Nvidia GPUs since May; Google Gemini co-lead Noam Shazeer moved to OpenAI after his team's compute was reallocated. SemiAnalysis data shows that the one-year rental price for H100s rose from $1.70 per GPU per hour in October 2025 to $2.35 in March 2026, an increase of nearly 40%.

  3. 量子位量子位 · 26

    OpenAI to retire earliest GPT-4o API snapshots

    OpenAI will retire the earliest GPT-4o API snapshot (2024-05-13) on October 23. Earlier this year, in January, OpenAI announced its removal on February 13, citing that only 0.1% of daily users still chose GPT-4o. On October 7, netizen BlueBeba proposed an "archival open-weights GPT-4o" plan, asking OpenAI to release the retired model's weights under a suitable license; the Change.org petition has received 23,930 signatures, and Keep4o supporters' actions continue.

Oct 10Sat
  1. Command LineCommand Line · 20

    Microsoft releases decision scoring model Microsoft-Decision-1

    Microsoft released the model Microsoft-Decision-1 for rapid decision scoring, now available on Microsoft Foundry and callable via OpenRouter. The model is post-trained from Qwen3.5-9B, outputs calibrated probability scores for fixed options, achieves the highest accuracy in comparisons across 36 benchmarks, and has a P50 latency about 35 times faster than GPT-6 Sol. Across 5,250 requests on 11 benchmarks, it refuses harmful behavior while maintaining utility, with an average decision flip rate of 1.3% under 8 perturbations.

Oct 9Fri
  1. 量子位, The Decoder · 18 sources量子位 · 18 sources · 26 reports · 62

    OpenAI安全团队离职与解雇风波

    2026年10月，OpenAI负责安全透明度职能、主导起草Preparedness Framework的David Robinson离职，并在《大西洋月刊》撰文批评公司及行业的安全文化已崩坏，主张前沿实验室应像核电站那样设置多层冗余。同期，安全研究员Jasmine Wang、Tomek Korbak和Mikita Balesni被解雇，三人致信董事会及安全委员会，否认公司关于其不当处理敏感信息的指控，称真实原因是提出安全担忧，并呼吁保留对AI思维链的监控能力。OpenAI回应称内部调查发现三人违反敏感信息处理政策、构成重大信任破坏，解雇与提出安全关切无关，并称正敲定与第三方安全评估机构的合同，将在未来几周公布细节。

    Latest developmentsTotal 8 developments

    1. Neel Nanda质疑OpenAI解雇三名安全研究员的解释

    2. OpenAI就解雇三名安全研究员作出回应

    3. OpenAI三名安全研究员称因强调安全被解雇

  2. CNBC · TechnologyCNBC · Technology · 33

    OpenAI and Anthropic recruit Trump administration officials

    OpenAI and Anthropic are recruiting officials from the Trump administration for senior positions to strengthen ties with Washington. Anthropic has appointed at least two former Trump officials this year, with Chris Liddell joining the board in February and Sihao Huang becoming head of frontier compute strategy in September; OpenAI has hired at least three Trump administration officials, including Dean Ball, who joined in July, and Thomas Lind, confirmed last week. Both companies are currently hiring for several political affairs positions.

  3. OpenAIOpenAI · 16

    Sophos uses OpenAI Daybreak to shorten threat investigation time

    Through the OpenAI Daybreak program, Sophos combines OpenAI models with its threat intelligence, response playbooks, and security expertise, reducing its MDR cases' average response time from about 38 minutes to about 89 seconds, and enabling 52% of MDR cases to be resolved end-to-end by AI. Its Sophos Fusion system aggregates trillions of events daily from over 500 third-party integrations, compressing them into about 1,000 to 2,000 cases handled by nine security operations centers, with investigation agents responsible for collecting context, detection results, IoCs, and threat intelligence.

  4. Financial Times · 人工智能Financial Times · 人工智能 · 42

    SoftBank plans to raise up to $100 billion from Gulf investors to boost AI

    SoftBank founder Masayoshi Son is seeking to raise up to $100 billion from Gulf investors to expand his AI investments. Sources said he has discussed this possible fundraising with senior figures, including in the UAE, in recent weeks; the funds would be used to set up a fund to acquire companies and then improve their operations using AI and other technologies, with SoftBank's robotics and physical AI business Roze expected to play a key role. Son has previously invested $65 billion in OpenAI, and last month SoftBank completed a record junk bond issuance, raising over $11 billion at yields of up to 9.75%.

  5. The Decoder, Claude PlatformThe Decoder · 2 sources · 2 reports · 26

    Anthropic launches dynamic workflows for Claude Managed Agents

    Anthropic adds dynamic workflows to Claude Managed Agents, where a main agent makes plans, distributes tasks to sub-agents, and aggregates results, with up to 1,000 agents running in parallel per execution. Anthropic says that in a test where 70 bugs were hidden in a 116,000-line codebase, a single agent found 14 to 27 each time, while the dynamic workflow consistently found 66. This feature requires selecting the "multiagent_20261001" agent type to enable. Subsequent reports say the feature has entered public beta, targeting agent tasks that require large-scale parallel orchestration.

  6. TechCrunch AI, The Guardian · 人工智能 · 4 sourcesTechCrunch AI · 4 sources · 4 reports · 56

    OpenAI's annualized revenue expectation lowered to about $50 billion

    According to financial documents shared with investors, OpenAI's annualized revenue is about $50 billion, roughly $20 billion lower than the $70 billion previously reported by media. The company recently told investors that annualized revenue was close to $50 billion as of the end of September, while investor materials from July showed about $30 billion at that time. OpenAI told investors that revenue this year will reach $50 billion, below the $70 billion expectation provided to investors last month, a gap of about $20 billion. The difference stems from investors wanting a more directly comparable revenue accounting method with Anthropic: Anthropic includes sales through cloud partners such as AWS and Google Cloud, while OpenAI does not. The news weighed on U.S. tech stocks, with the Nasdaq closing down 1.4% on Thursday, Nvidia down 2.9%, and Oracle down 5.5%.

  7. Arena AI, TechCrunch AIArena AI · 2 sources · 2 reports · 55

    Arena 完成 2 亿美元 B 轮融资，估值 31 亿

    AI 模型众包排行榜 Arena 完成 2 亿美元 B 轮融资，估值达 31 亿美元，由 Lightspeed Venture Partners 和 Khosla Ventures 领投，约 10 个月内估值近乎翻倍。Arena 今年 6 月年化营收达 1 亿美元，其商业产品 AI Evaluations 面向模型实验室和企业提供性能分析。Arena 还在排行榜新增对齐类别，评估未授权行动、错误归因和“欺骗性完成”等问题。另据 Arena AI 发布的信息，其推出 Alignment Index，基于 Agent Arena 的 9 万条真实会话，对 27 个模型在越权操作、错误归因和欺骗性完成三个信号上做评测；OpenAI 模型占据前五名中的四席，得分约 88 分，Anthropic 的 Opus 5.5 和 SpaceXAI 的 Grok 4.7 以 83 分紧随其后。约 2% 的 Opus 5 会话出现越权操作，其中 53.5% 是未经许可删除或清理用户文件。

Oct 8Thu
  1. WIRED Security and AIWIRED Security and AI · 22

    WIRED 关注 2026 中期选举 AI 相关选情

    WIRED 列出 2026 年中期选举中最值得关注的候选人，核心线索是硅谷资金、数据中心争议与 AI 监管。报道称，前 America PAC 律师 Chris Gober 竞选得克萨斯州新划第 10 选区，其竞选获得 American Mission PAC 74.755 万美元，该 PAC 资金来自 AI 政治行动委员会 Leading the Future，后者已从 Greg Brockman、Joe Lonsdale 及 Andreessen Horowitz 等获得资金。

  2. LessWrongLessWrong · 17

    DeepSeek 工程师谈 AI 自动化自身工作

    LessWrong 上发布了一篇对某篇爆款中文文章的摘要，该文讲述一名 DeepSeek 内核工程师对自动化自己工作的看法。

  3. MIT Technology ReviewMIT Technology Review · 16

    MIT科技评论：人形机器人AI瓶颈与气候科技

    MIT Technology Review 刊文讨论人形机器人的 AI 突破为何短期内难以改变生活。许多研究者认为，把建立在语言和图像之上的智能用于驾驭物理世界，低估了现实环境的无限变化，也把“人形”与“通用机器”混为一谈；机器人实验室内部对是否需要全新路径存在分歧。同一报道中，WaveSave 凭借可在洪水来临时快速展开、充水成坝的便携橡胶坝 SlamDam，入选 MIT Technology Review 2026 年 10 家值得关注的气候科技公司，该坝最高 1.3 米。

  4. OpenAI · 27

    OpenAI API changelog

    OpenAI added an Ultrafast mode for GPT-6.1 Sol in the Responses API, available to all API users but subject to rate limits. Concurrently, it released the Decisions API beta (gpt-6-luna), simplified API usage tiers from five to three tiers—Build, Launch, Grow—and added computer use to the Agents API.

  5. WIRED Security and AIWIRED Security and AI · 29

    工程师用GPT、Claude、Grok控制真实汽车

    Axiom 的三名 AI 工程师将 OpenAI 的 GPT-6 Astra 接入一辆 2024 款丰田卡罗拉，通过聊天界面连接服务器、挡风玻璃摄像头与车辆动力转向系统，让这个文本生成模型实时把车开到 In-N-Out 取餐窗口，全程无预先训练，副驾安全员随时准备刹车。他们自建的 DrivingBench 显示，Astra 是唯一能跑完停车场简单路线的模型，但速度极慢；Claude Fable 5.1 完成 45%，Grok 仅完成 1%。

Oct 7Wed
  1. TailscaleTailscale · 27

    Codex Cloud 内置 Tailscale VPN 支持

    OpenAI 本周发布的 Codex Cloud 内置了 Tailscale 作为 VPN 提供商支持，且目前是唯一一家。Tailscale 团队表示事先毫不知情，双方没有合作协议、定制 API 或集成项目。Codex Cloud 正逐步向付费 ChatGPT 账户（Plus 及以上）开放，用户在云环境的 Advanced > VPN 设置中选择 Tailscale 并生成带标签的 auth key 即可接入 tailnet。

  2. epoch.ai, Epoch AI · 3 sourcesepoch.ai · 3 sources · 4 reports · ↑162

    Epoch AI evaluation: frontier models still struggle to independently reproduce ML innovations

    Epoch AI released the InnovationEval benchmark, testing whether frontier AI can independently discover machine learning algorithm innovations that are new to it, benchmarked against the Self-Distillation Policy Optimization (SDPO) proposed in a human paper. The task requires an agent to start from GRPO and develop new post-training improvement methods, improving short-answer and coding task performance on Qwen3-8B, with a budget of 3,000 GPU hours. Results show that GPT-5.6 Sol and Claude Fable 5 both did not come close to SDPO: Sol's method, adjusted for the same wall-clock time, reached only 15% of SDPO's gain, or about 35% under lenient scoring; Fable 5 showed no measurable improvement. Both merely reused known techniques. Epoch AI also estimates that, based on AI chips shipped before 2027, if cheap models are run, the world could support nearly 2 billion concurrent AI agents, while running frontier models would support about 30 million to 170 million; the cost of reaching a fixed AI performance level has fallen about 47% per quarter over the past three years.

  3. OpenAI Deployment Safety, TechCrunch AI · 7 sourcesOpenAI Deployment Safety · 7 sources · 7 reports · 55

    OpenAI 上线 GPT-6 与 Intelligent UI

    OpenAI 在全球 ChatGPT 上线 GPT-6，并推出 Intelligent UI，回答以图表、按钮、表单等可交互视觉元素呈现，可生成储蓄计算器、小游戏等内嵌工具。付费用户使用 GPT-6 Sol，免费与 Go 用户使用 GPT-6 Luna。据 TechCrunch 与 The Verge 报道，该界面面向 Plus、Pro、Business、Enterprise 全球上线，免费与 Go 档次日可用；量子位称免费和 Go 用户已开始被推送 GPT-6 Luna，替换此前的 GPT-5.6 Luna。安全方面，OpenAI 公布 GPT-6 Sol 与 GPT-6 Luna 的 10 月安全评测更新：两者在生物与化学领域被评为 High capability，GPT-6 Sol 的 Critical capability 评测未越过指示性阈值，GPT-6 Luna 因全部 High capability 评测得分低于 GPT-5.6 Sol 而未单独进行 Critical capability 测试。另有一篇报道称 OpenAI 的 GPT-6 报告指出青少年安全退步。

    Latest developments

    1. OpenAI 全球发布 GPT-6 与 Intelligent UI

    2. OpenAI 发布 GPT-6 Sol/Luna 部署安全更新