今日重点

2 项需处理

08.27 2026.08.27 · 01:25 · 北京时间
来源更新正常
进行中 · Kimi · 下线

kimi-k2 下线

kimi-k2 系列模型已于 2026 年 5 月 25 日下线 ,不再维护和支持。请使用最新模型 kimi-k3 ,以获得持续支持和更强推理能力。

查看详情
级别
公开证据
1
更新时间
00:17
Kimi

kimi-thinking-preview 下线

kimi-thinking-preview 已于 2025 年 11 月 11 日下线 ,不再维护和支持。建议直接升级至最新模型 kimi-k3 ,以获得思考能力。

暂不计算等待更多证据

其他变化

按行动优先级排序
Kimi

kimi-latest 下线

kimi-latest 已于 2026 年 1 月 28 日下线 ,将不再维护和支持。请直接使用 Kimi 最新模型 kimi-k3 ,以获得持续支持和更强推理能力。

暂不计算等待更多证据
Kimi

kimi-k2.5, moonshot-v1 下线

Kimi K3 发布后, kimi-k2.5 和 moonshot-v1 系列模型已停止向新注册用户开放(全平台正式下线时间为 8 月 31 日),请尽快切换至新模型。

暂不计算等待更多证据
OpenAI

Elevated error rates

All impacted services have now fully recovered.

已解决事件已结束
智谱 GLM

GLM-5.3-Flash 原生多模态模型上线

原生融入视觉能力,使模型能够主动观察界面、渲染与交互反馈并持续迭代改进,实现代码、浏览器与图形界面的协同闭环。 极致高效混合架构:采用线性注意力与稀疏注意力混合架构(总参 320B,激活 18B),计算量与 KV 缓存较 GLM-5.3 大幅降低。 超越 Coding 的专业工作伙伴:拓展至 Office 文档与金融研究工作流,自主拆解复杂目标、调用工具、检查优化输出。

暂不计算等待更多证据
智谱 GLM

GLM-5.3-Flash 轻量高速视觉模型上线

延续 GLM-5.3 基座的视觉理解与智能体能力,轻量高速,兼顾性能与效率 支持最长 1M 上下文处理与原生 Function Calling,始终开启思考模式(不支持 disabled) 兼顾性能与效率,适合高并发、低延迟的多模态业务场景

暂不计算等待更多证据
OpenAI

API customers can now select regional processing for an individual request by using a prefixed domain with an API key from a project having Global geography.

API customers can now select regional processing for an individual request by using a prefixed domain with an API key from a project having Global geography. Existing eligibility, data retention control, endpoint, and model support requirements continue to apply. Learn more in the data controls guide .

暂不计算等待更多证据
OpenAI

Released the Prompt Caching dashboard on the OpenAI API platform.

Released the Prompt Caching dashboard on the OpenAI API platform. Track your cache hit rate over time, cache reads per write, and the breakdown of cache-read, cache-write, and uncached tokens to understand your caching efficiency and identify opportunities to improve. Filter metrics by model and service tier.

暂不计算等待更多证据
OpenAI

Transparent backgrounds are now available in preview for gpt-image-2 and gpt-image-2-2026-04-21 in the Images API and the Responses API image generation tool.

Transparent backgrounds are now available in preview for gpt-image-2 and gpt-image-2-2026-04-21 in the Images API and the Responses API image generation tool. Set background to transparent and use png or webp output; jpeg does not support transparent backgrounds. Learn more in the image generation guide .

暂不计算等待更多证据
智谱 GLM

GLM-5.3 新一代旗舰模型上线

更强的编程能力:GLM-5.3 的编程能力大幅提升,在智谱内部 Z.ai Code Bench 上较 GLM-5.2 提升了 50%,在包括 Terminal Bench 3.0 等公开基准测试中达到开源模型 SOTA 水平 涌现的网络安全能力:在白盒代码审查与漏洞发现持平 Mythos 5,已联合多个中国安全团队测试并累计发现 2436 个漏洞(1097 个中高危)

暂不计算等待更多证据
OpenAI

Daybreak models are now available on AWS

OpenAI and AWS are making Daybreak cybersecurity capabilities available through Amazon Bedrock to support enterprise security workflows.

暂不计算等待更多证据
OpenAI

Daybreak now offers two access tiers for approved defenders: Daybreak Blue and Daybreak Red.

Daybreak now offers two access tiers for approved defenders: Daybreak Blue and Daybreak Red. Use them to move from security findings to validated fixes in explicitly authorized engagements. Start with Daybreak Blue for most defensive security work. It provides access to general-purpose models such as GPT-5.6 Sol for vulnerability discovery, secure code review, detection engineering, incident response, malware analysis, and patch validation. Read more here . Daybreak Red provides separately approved access to purpose-trained models such as GPT-5.6 Cyber for authorized vulnerability reproduction, exploit validation, penetration testing, red teaming, and complex system analysis. These models require separate approval and provisioning. You can apply to join the Daybreak program here . More details on pricing here .

暂不计算等待更多证据
OpenAI

Starting July 30, GPT-5.6 Luna costs 80% less, while GPT-5.6 Terra costs 20% less.

Starting July 30, GPT-5.6 Luna costs 80% less, while GPT-5.6 Terra costs 20% less. See pricing details . We're also introducing Fast mode in the API, which replaces our Priority Processing offering. For GPT-5.6 Sol, Fast mode now delivers up to 2.5× faster speeds than standard processing at twice the price. This change is backward compatible: requests tagged priority will automatically use Fast mode.

暂不计算等待更多证据
OpenAI

Released the official OpenAI Terraform provider for managing OpenAI API Platform resources as infrastructure as code.

Released the official OpenAI Terraform provider for managing OpenAI API Platform resources as infrastructure as code. Provision and manage projects, users, groups, roles, access assignments, service accounts, certificates, invitations, and project-level rate limits. Use standard Terraform workflows to review and apply changes, import existing resources, and detect and reconcile configuration drift. Install the provider from the Terraform Registry .

暂不计算等待更多证据
OpenAI

Released GPT Transcribe for accurate file transcription and final transcripts of committed Realtime turns, along with GPT Live Transcribe for low-latency streaming transcription.

Released GPT Transcribe for accurate file transcription and final transcripts of committed Realtime turns, along with GPT Live Transcribe for low-latency streaming transcription. Both models support free-form transcription context, keyword hints, and multiple expected input languages. Compare supported outputs and workflows in the transcription guide .

暂不计算等待更多证据
Kimi

Kimi 全球大使计划现已开启

Kimi 面向全球招募在真实场景中深度使用 Kimi、并愿意将经验与洞察分享给更多人的先行者。

暂不计算等待更多证据
OpenAI

Launching Health in ChatGPT

Health in ChatGPT now lets eligible U.S. users securely connect medical records and Apple Health to get more personalized insights and better understand their health.

暂不计算等待更多证据
OpenAI

Introducing OpenAI Presence

Introducing OpenAI Presence, a proven enterprise AI agent platform that helps organizations deploy trusted voice and chat agents for customer and internal workflows.

暂不计算等待更多证据
OpenAI

Added hard spend limits for organizations and projects on the OpenAI API platform.

Added hard spend limits for organizations and projects on the OpenAI API platform. Set a monthly cap that causes affected API requests to return a 429 error when tracked spend reaches the limit. Use spend alerts for notification before traffic is interrupted. Read more in the spend limits guide .

暂不计算等待更多证据
Kimi

Kimi K3:智能的新前沿

Kimi K3 正式发布:2.8 万亿参数、原生视觉理解、100 万 token 上下文,全球首个开源的 3 万亿级别模型。

暂不计算等待更多证据
OpenAI

Released the GPT-5.6 model family , including GPT-5.6 Sol for frontier capability, GPT-5.6 Terra for a balance of intelligence and cost, and GPT-5.6 Luna for efficient, high-volume workloads.

Released the GPT-5.6 model family , including GPT-5.6 Sol for frontier capability, GPT-5.6 Terra for a balance of intelligence and cost, and GPT-5.6 Luna for efficient, high-volume workloads. The gpt-5.6 alias routes requests to gpt-5.6-sol . GPT-5.6 adds Programmatic Tool Calling , explicit prompt caching controls , persisted reasoning, max reasoning effort, and Pro mode , and Multi-agent orchestration in beta for the Responses API . GPT-5.6 also accepts images at their original dimensions with original or auto image detail.

暂不计算等待更多证据
OpenAI

Introducing GPT-Live

A new generation of voice models for natural human-AI interaction, now powering ChatGPT Voice.

暂不计算等待更多证据
OpenAI

Released the Safety Usage Dashboard on the OpenAI API platform.

Released the Safety Usage Dashboard on the OpenAI API platform. The Safety dashboard shows blocked Responses requests based on safety_identifier values sent on requests to identify end users. Visit the Safety dashboard .

暂不计算等待更多证据
智谱 GLM

GLM-5.2 新一代旗舰模型上线

支持 1M 无损上下文,长程任务能力显著提升,减少复杂任务中的上下文漂移与目标遗忘 Coding 与长程任务评测达到开源 SOTA,在复杂系统工程、深度调试中表现更稳 真实开发体感显著提升,项目级上下文承载、工程规范遵循与多端开发更可靠

暂不计算等待更多证据
OpenAI

Introducing the OpenAI Partner Network

OpenAI launches the Partner Network, investing $150M to help global partners accelerate enterprise AI adoption, deployment, and transformation.

暂不计算等待更多证据
OpenAI

Web search can now return image results alongside regular text results.

Web search can now return image results alongside regular text results. Use image search when your application needs current or web-grounded visuals, such as product photos, landmarks, places, events, or visual references. Read more in the web search guide .

暂不计算等待更多证据
OpenAI

Added moderation scores to the Responses API and Chat Completions API.

Added moderation scores to the Responses API and Chat Completions API. Pass a moderation object in a generation request to receive moderation results for both the model input and generated output in the same response. Learn more in the Moderation guide .

暂不计算等待更多证据
OpenAI

Introducing new capabilities to GPT-Rosalind

GPT-Rosalind advances life sciences research with enhanced biological reasoning, medicinal chemistry expertise, genomics analysis, and experimental workflow capabilities.

暂不计算等待更多证据
OpenAI

Starting June 2, 2026, eligible container sessions will be billed per minute with a 5-minute minimum, instead of being billed at the full 20-minute session rate.

Starting June 2, 2026, eligible container sessions will be billed per minute with a 5-minute minimum, instead of being billed at the full 20-minute session rate. The underlying per-minute rate will remain the same. This update is intended to make billing more granular for shorter sessions and will lower effective cost for customers. You can find current built-in tool pricing in our API pricing docs .

暂不计算等待更多证据
OpenAI

OpenAI frontier models and Codex are now available on AWS

OpenAI frontier models and Codex are now generally available on AWS, giving enterprises a new path to build with OpenAI through the AWS environments, controls, and procurement workflows they already use. Customers can get started with OpenAI on AWS and move faster from evaluation to production.

暂不计算等待更多证据
智谱 GLM

GLM Coding Plan 团队版上线

面向企业与开发团队的自助订阅方案上线,延续个人版高额模型用量,并兼容全球主流编码工具 支持席位、权限、用量与预算统一管理,帮助团队实现可追踪、可控制的 AI 编程协作 提供集中账单、统一开票与企业级数据安全保障,默认不将代码、提示词和对话内容用于模型训练 高级版支持首发接入最新旗舰模型及高峰期资源优先保障,带来更稳定高效的团队开发体验

暂不计算等待更多证据
OpenAI

Released workload identity federation .

Released workload identity federation . Trusted workloads can exchange externally issued identity tokens for short-lived OpenAI access tokens without storing long-lived API keys.

暂不计算等待更多证据
OpenAI

Released Secure MCP Tunnel for enterprise customers.

Released Secure MCP Tunnel for enterprise customers. Secure MCP Tunnel lets supported OpenAI products including ChatGPT web, Codex, Responses API, and AgentKit connect to private or on-prem MCP servers through a customer-hosted tunnel-client without exposing those servers to the public internet.

暂不计算等待更多证据
OpenAI

Deprecated DALL·E model snapshots and the Realtime API Beta.

Deprecated DALL·E model snapshots and the Realtime API Beta. DALL·E model snapshots dall-e-2 and dall-e-3 were deprecated and removed from the API on May 12, 2026. We recommend using gpt-image-2 , gpt-image-1 , or gpt-image-1-mini instead. The Realtime API Beta was deprecated and removed from the API on May 12, 2026. If you are still using the beta interface, migrate to the released Realtime API. See the migration guide and the full deprecations page .

暂不计算等待更多证据
OpenAI

Released GPT-Realtime-2 , a new realtime voice model with configurable reasoning for speech-to-speech agents, along with GPT-Realtime-Translate for streaming speech translation and GPT-Realtime-Whisper for streaming speech-to-text.

Released GPT-Realtime-2 , a new realtime voice model with configurable reasoning for speech-to-speech agents, along with GPT-Realtime-Translate for streaming speech translation and GPT-Realtime-Whisper for streaming speech-to-text. Updated the Realtime and audio guide , added a dedicated Realtime translation guide , refreshed Realtime transcription for streaming transcripts, and moved realtime prompting guidance into Using realtime models .

暂不计算等待更多证据
OpenAI

Released the OpenAI Developers plugin for Codex .

Released the OpenAI Developers plugin for Codex . This helps you build AI applications and agents in Codex with OpenAI Platform access and OpenAI API setup guidance.

暂不计算等待更多证据
OpenAI

Introducing Advanced Account Security

Introducing Advanced Account Security: phishing-resistant login, stronger recovery, and enhanced protections to safeguard sensitive data and prevent account takeover.

暂不计算等待更多证据
智谱 GLM

GLM-5.1 新一代旗舰模型上线

Coding 能力大大增强,长程任务(Long Horizon Task)显著提升,支持一次任务中独立、持续工作长达 8 小时,实现从规划、执行到交付的完整闭环 在自主规划、持续执行、问题修复与策略迭代上展现更强的工程智能,能够完成更长链路的复杂任务闭环 综合能力全面对齐 Claude Opus 4.6,成为首个在综合能力上实现全面对齐的中国模型,并跻身全球开源模型前列 通过 multi-turn SFT、RL 与过程质量评估体系,进一步强化长任务中的稳定性、一致性与 tool use 能力

暂不计算等待更多证据
智谱 GLM

GLM-5V-Turbo 多模态 Coding 基座模型上线

兼顾视觉理解与 Coding 能力,在更小参数量下实现更优的性能表现,多模态任务处理更高效 细粒度理解、几何感知与空间理解能力进一步增强,复杂视觉推理更准确 强化 GUI Agent、Coding Agent 等复杂任务表现,更适合“看懂环境—规划动作—执行任务”的长流程场景 多模态工具链进一步扩展,在文本工具基础上新增支持画框、截图、读网页(含图片识别)等多模态 Tools

暂不计算等待更多证据
智谱 GLM

GLM-5-Turbo 龙虾增强基座模型上线

面向 OpenClaw 龙虾场景深度优化的基座模型 强化了对外部工具与各类Skills的调用能力,在多步任务中更稳定、更可靠 复杂指令拆解更强,能够精准识别目标、规划步骤,并支持多智能体之间的协同分工 能够更好理解时间维度上的要求,在复杂长任务中保持执行连续性 针对数据吞吐量大、逻辑链条长的龙虾任务,进一步提升了执行效率与响应稳定性

暂不计算等待更多证据
智谱 GLM

GLM-5 新一代旗舰模型上线

专为复杂系统工程与长程 Agent 任务设计,实现从代码到工程的范式跃迁 后端架构设计、复杂算法实现及顽固 Bug 修复上展现出卓越的深度推理能力 在代码逻辑密度和系统工程能力上直接对标 Claude Opus 4.5 首次集成 DeepSeek Sparse Attention,在维持长文本效果无损的同时,提升 Token Efficiency

暂不计算等待更多证据
智谱 GLM

GLM-OCR 图文解析模型上线

采用自研 CogViT 与 GLM-0.5B 的编码器-解码器设计,连接层实现高效跨模态对齐 基于数十亿图文对的 CLIP 预训练,具备强大的视觉语义与关键 Token 提取能力 模型小、速度快,在手写体、表格、印章、竖排等复杂场景中表现稳定

暂不计算等待更多证据
智谱 GLM

GLM-4.7-Flash 免费模型上线

轻量参数规模下实现了高效的 Coding 能力,任务理解与代码生成能力处于同类模型的较高水平 通用能力同级别最优,在写作、翻译、推理、角色扮演、长文本与审美等核心场景下兼顾质量与响应速度 为 GLM-4.7 的免费版本,针对高频调用场景进行优化,显著降低使用门槛与成本

暂不计算等待更多证据
智谱 GLM

GLM-Image 图像生成模型上线

首个在国产芯片上完成全流程训练的 SOTA 多模态模型 采用「自回归理解 + 扩散解码」混合架构,模型能读懂指令、补全细节 知识密集型场景全面增强,文字渲染更稳更准(汉字尤其出色)

暂不计算等待更多证据
智谱 GLM

GLM-4.7 基座模型上线

Coding 能力全面提升,代码生成更稳、更完整,长代码与工程级场景下的一次性交付能力显著增强 Agentic Coding 能力升级,支持以任务为中心的端到端开发 前端与视觉代码理解增强,生成页面在布局、交互与审美上更接近可展示、可直接使用的水平 通用对话与内容生成更可靠,复杂问题拆解更清晰,简单问题回应更直接,整体表达更自然、高效

暂不计算等待更多证据
智谱 GLM

GLM-TTS-Clone 音色克隆模型上线

只需录制约 3 秒清晰语音,即可生成专属音色 支持普通话及轻口音,更好复刻节奏、停顿、语气词等个性化表达 无论旁白、客服、剧情配音、教育讲解,均可保持音色统一与情感一致性 可与 GLM-TTS 联动,通过音色 ID 进行批量生产

暂不计算等待更多证据
智谱 GLM

AutoGLM-Phone AI 手机智能助理框架上线

支持用自然语言自动完成 App 操作任务 具备界面识别、意图规划与设备执行的端到端处理能力,无需人工点击或复杂配置 已适配 50+ 主流中文应用场景,覆盖购物、出行、外卖、影音、资讯等高频任务 支持完整操作指令集,包括启动 App、输入文字、滑动、点击、回退、长按等,实现细粒度交互控制

暂不计算等待更多证据
智谱 GLM

GLM-TTS 语音合成模型上线

在架构上采用两阶段生成,在自然度、情绪表达与语调连贯性上实现全面升级 支持流式与非流式接口,首帧响应低于 400ms,实现低延迟的交互式体验 强化学习(GRPO)优化显著降低字错误率,并在开源评测中取得 SOTA 情感表达表现 支持灵活控制语速、音量与风格,满足客服、阅读、文旅、智能硬件等多场景需求

暂不计算等待更多证据
智谱 GLM

GLM-ASR-2512 语音识别模型上线

行业出色语音识别性能,最新评测字符错误率(CER)仅 0.0717 支持多语言与方言识别,覆盖普通话、粤语、四川话、美式/英式英语及数十种全球常用语言 原生支持自定义词典,可快速导入专业术语、人名地名与项目代号,显著提升垂直行业识别精度

暂不计算等待更多证据
智谱 GLM

GLM-4.6V 视觉推理模型上线

20+ 主流多模态评测基准全面验证,均取得 SOTA 成绩 原生支持工具调用,具备强大的图文混排创作能力,并能处理识图购物等复杂的视觉任务 视觉上下文窗口扩展至 128k,可单次处理约 150 页的复杂文档、200 页 PPT 或一小时视频 针对前端开发场景深度调优,支持“截图即代码”,大幅提升 GLM Coding Plan 的开发与调试效率

暂不计算等待更多证据
智谱 GLM

GLM-4.6 基座模型上线

在公开基准与真实编程任务中展现出更强的代码能力 上下文窗口扩展至200K,提升处理长代码与复杂智能体任务的能力 推理过程进一步优化,支持在推理过程中调用工具,提升智能体任务执行的灵活性与效率 加强了模型在工具调用和智能体框架下的表现,使其在多场景应用中更加高效

暂不计算等待更多证据
智谱 GLM

GLM-4.5V 视觉推理模型上线

100B 级别开源视觉推理模型 SOTA,比 GLM-4.1V-Thinking “更大更强” 覆盖从视频理解、前端复刻、视觉定位、图像识别与推理,到复杂文档解析与 GUI Agent 等多场景的视觉任务 新增“思考模式”开关,可灵活选择快速响应或深度推理

暂不计算等待更多证据
智谱 GLM

GLM-4.5 系列基座模型上线

SOTA 级原生智能体大模型 参数效率翻倍,API 价格仅为 Claude 的1/10,极速版速度超 100tokens/秒 实测 Agentic Coding 表现优异,支持一键兼容 Claude Code 框架

暂不计算等待更多证据
智谱 GLM

CogVideoX-3 视频生成模型上线

新升级视频生成大模型,支持文生、图生视频 新增首尾帧生成功能 画面清晰度主观感受显著提升 主体大幅度运动自然流畅 提升了高清现实及 3D 风格场景表现

暂不计算等待更多证据
智谱 GLM

GLM-4.1V-Thinking 系列视觉推理模型上线

定位优势:通用视觉模型 核心能力:具备强大的多模态理解和推理能力 任务表现:在视频理解、图像问答、图表解读、图形界面操作等多任务项均达到新SOTA 能力特点:不止看得懂(基础视觉理解),更能想得透(深度推理能力)

暂不计算等待更多证据
智谱 GLM

接入两个 Vidu 热门视频生成模型

聚焦高质量视频创作 固定输出 5 秒、24 帧、1080P 规格内容 凭借对清晰度的深度优化,画质质感大幅跃升 写实风格逼近真实场景,2D 动画画风精准保持 首尾帧转场更加丝滑 适用于影视、广告、动漫短剧等高要求创作场景 平衡速度、质量与成本 主攻图生视频、首尾帧功能 支持 4 秒时长下 720P分辨率输出 画面稳定可控适配电商等场景 首尾帧语义理解与多参考图一致性增强 是泛娱乐、互联网、动漫短剧、广告量产的高效工具

暂不计算等待更多证据
智谱 GLM

新增高识别精度、强抗噪能力的语音模型

能够基于上下文理解将音频转录为符合语言习惯的文本 显著提升输出结果的流畅性和可读性 在噪音环境中较当前模型有明显较好的表现 不会被非语言类噪声干扰 支持中文、英语以及各地方方言(东北官话、胶辽官话、北京官话、冀鲁官话、中原官话、江淮官话、兰银官话和西南官话) 上新限时免费中~

暂不计算等待更多证据
智谱 GLM

一次新增 2 个基座模型、3 个推理模型

在工程代码、Artifacts 生成、函数调用、搜索问答及报告撰写等任务上均表现出色 性能比肩 GPT-4o、DeepSeek-V3-0324 等大尺寸模型水平 免费使用 在通用任务上依然表现出色 适合轻量化任务 性能优异的推理模型 速度最快可达 200 tokens/秒(比常规快 8 倍) 价格仅为 DeepSeek-R1 的 1/30 适合高频调用场景 免费使用 进一步降低模型使用门槛

暂不计算等待更多证据
智谱 GLM

一站式 AI 搜索工具全家桶更新升级

Web Search API: 支持直接获取结构化搜索结果(标题/摘要/链接等),提供多搜索引擎支持(智谱自研/bing/搜狗/夸克/Jina AI)。 Chat Search: 将 Web Search API 的搜索结果融入大模型,生成智能回答,并标注网页结果来源。 Search Agent: 根据用户的 query,分析搜索意图,进行多个 query 拆解,将搜索结果融入大模型,提供全面、有深度的回答。

暂不计算等待更多证据