GPT Intelligence
AI model intelligence from AIYXL.
Developed By
Related Applications
Knowledge Areas
Related Articles
7.9 国际发售部分获得超1000亿港元机构订单,覆盖15个国家和地区的主权基金与长线基金,国际发售超额认购约44倍
1.15 有分析认为,在全球大模型生态日益开放的背景下,基于优质开源基座进行本地化优化与商业化落地,本身就是一种合理且高效的创新路径
12.27 我们发布了中文大模型基准测评2024上半年报告,在AI大模型发展的巨大浪潮中,通过多维度综合性测评,对国内外大模型发展现状进行观察与思考。
10.17基于Sonnet4.5的代理可将简单任务交由Haiku4.5子代理处理,从而降低多步骤编程或市场研究类工作流的推理成本
11.20 该技术允许模型在接近上下文窗口极限时,智能地识别保留核心逻辑与状态信息,同时丢弃冗余历史,从而实现跨多个上下文窗口的连贯持久推理
9.4 当地时间9月3日,OpenAI正式发布新一代旗舰模型GPT-6 Astra,这是该公司迄今为止最强大、对齐程度最高的模型。OpenAI总裁格雷格·布罗克曼在发布会最后留下一句话:“欢迎来到AGI时代。”
5.14 据NVIDIA称,当AzureAI智能体服务与开源工具包NVIDIAAgentIQ配对时,开发人员现在可以实时分析和优化AI代理团队,以减少延迟提高准确性并降低计算成本
5.2 产品价格语音AgentAPI(语音到语音)3美元小时(0.05美元分钟)文本转语音TTS4.20美元百万字符语音转写STT(实时流式)0.20美元小时语音转写STT(批量)0.10美元小时
7.30 On July 29, 1,134 employees from leading companies including OpenAI, Anthropic, Google DeepMind, and Meta signed a joint open letter titled "Pacing the Frontier," calling on the U.S. government to take the lead in promoting international cooperation and establishing a globally coordinated governance framework that can actively regulate the pace of frontier AI development.
7.9 On July 9, OpenAI officially released the GPT-Live series of voice models, including GPT-Live-1 and GPT-Live-1 mini, rolling out globally to ChatGPT users. GPT-Live-1 is available to Plus, Pro, and Go subscribers, while GPT-Live-1 mini is for Free users, with API access planned soon.
5.24 The year 2026 is destined to be a blockbuster year for IPOs in global tech history. SpaceX formally submitted its S-1 registration statement to the SEC on May 20, with ticker symbol "SPCX," targeting a June 11–12 pricing and a late-June listing, aiming to raise US$75–80 billion at a US$1.75 trillion valuation. OpenAI confidentially filed its IPO documents on May 22, targeting a September listing at a valuation of US$852 billion to US$1 trillion. Anthropic, after completing its US$90 billion
2.28 实测显示,模型可根据UI设计图直接生成前端代码,或根据数据库架构图生成SQL查询,在"设计-开发"一体化场景中展现强大潜力
9.2 On September 1, Anthropic announced the release of Claude Fable 5.1 and Claude Mythos 5.1. The two models share the exact same foundation model but differ in safety guardrails: Fable 5.1 is open to all users, while Mythos 5.1 is available only through the Trusted Access Program to vetted cybersecurity and life sciences organizations.
4.15 三款新模型:GPT-4.1、GPT-4.1mini、GPT-4.1nano 集体亮相,编程、长文本、指令理解能力全面升级!
10.1 这款拥有320亿参数的AI系统,全球首个将世界模型系统性引入代码生成任务的LLM,与传统代码模型直接输出代码不同,CWM在生成代码前会在沙箱环境中模拟和推演,预测某一行代码运行后的后果
12.10 该工具采用Apache2.0协议开源,支持自定义编程和主题切换,个人和商业场景均可免费使用
5.22 公司在发布证明的同时附上了一份多位顶尖数学家联署的伴随论文,包括菲尔兹奖得主TimothyGowers普林斯顿大学数学家NogaAlon多伦多大学数论学家ArulShankar以及维护埃尔多斯问题网站的数学家ThomasBloom
9.8 训练使用了欧盟委员会大型人工智能大奖挑战赛提供的200万GPU小时的计算资源,这些计算资源是在欧洲的超级计算机LUMI芬兰和JUPITER上完成的
2.17 与此前“t语言模型+视觉模块”t的两段式架构不同,Qwen3.5直接把视觉感知和语言推理塞进了同一个训练框架,在文本和视觉的混合数据上预训练,实现了视觉理解能力的质变
3.12 这些工具是OpenAI新响应API的一部分,它允许企业开发自定义AI智能体,这些智能体可以执行网络搜索扫描公司文件和导航网站,就像OpenAI的Operator产品一样