PatrocinadoChatgpt/Claude/Grok 会员24h在线自助充值 安全快速 稳定保障
奥创(Ultra)商城会员充值中心
PatrocinadoChatgpt/Claude/Grok 会员24h在线自助充值 安全快速 稳定保障
奥创(Ultra)商城会员充值中心
Patrocinado本页公开展示实时样本、模型覆盖与响应时间。先用 5 元按量跑自己的 Claude Code/Codex 任务,成功后再决定是否长期使用。
ACS Gateway
Patrocinado【腾讯云】WorkBuddy 全场景 AI 办公工作台,团队旗舰版低至198元/月
腾讯云
Patrocinado人人可用的AI应用创作平台 自然语言交互,全栈应用搭建,伟大创意快速实现;个人版新用户注册送5000积分
阿里云Atualizações com fontes verificadas sobre modelos de IA, APIs, investigação, segurança e regulação. Consulte as evidências e acompanhe alterações de preços e modelos.
O conteúdo está em inglês; a navegação e os filtros mantêm o idioma selecionado.
Z.ai has announced GLM-5.3-Flash, a 320-billion-parameter mixture-of-experts model with approximately 18 billion active parameters per token, native multimodal support, and a context window of up to 1,048,576 tokens.
Google DeepMind and Google Research have introduced WeatherNext 3, a global forecasting model that uses real-time observations and raw satellite data to produce hourly, high-resolution predictions.
Gemini 3.8 Flash Cyber is designed for security defenders that need fast vulnerability analysis and automated code fixes. Google is initially providing access only to trusted defenders through its Fairwind Program.
Google has introduced agentic video understanding for its latest Gemini models. The capability dynamically searches long-form video and, in company-reported benchmarks, used up to 88% fewer tokens while improving analysis accuracy.
Gemini Omni 1.1 Flash adds scene extension, first- and last-frame controls, faster low-resolution drafts, and high-resolution video output for developers.
Z.ai says every current GLM Coding Plan subscriber will receive a Reset Card that can replenish both weekly and five-hour usage quotas.
Z.ai has released GLM-5.3 as an open-weight model aimed at agentic coding and cyber defense. Its weights are available on Hugging Face alongside a technical blog.
SGLang and RadixArk have released Miles v0.1, an open-source reinforcement-learning framework for post-training large language and multimodal models. It focuses on debugging reliability, hardware efficiency, and scalable execution.
OJO has introduced a design workspace where users can assemble specialized AI agents and move product ideas from strategy and PRDs to interactive prototypes and launch-ready designs.
Moonshot AI’s tutorial demonstrates how Kimi Work can help build a live investor dashboard, update spreadsheet-based financial models, and process research reports in batches.
ColaMD 1.9.0 adds DOCX, image, and improved PDF export, along with an outline view, autosave, automatic updates, and changelog notifications. Windows startup performance has also been improved.
Developer Guizang has built a UI and interaction enhancement plugin for DeepSeek Harness, incorporating assets developed for Codepilot. The project was later released as Pilot Harness, an out-of-the-box client built around DeepSeek Harness plugins.
Mole is a macOS utility that combines cleanup, app removal, disk-space analysis, optimization, and system monitoring. Its command-line edition remains free and open source, while a native Mac app provides a graphical experience.
The post recounts a comment seen while watching an interview with Sam Altman. Available evidence confirms related X posts and a Codex discussion, but does not establish a separate product-news claim.
MiniMax says it demonstrated its H3 model at a creator-focused event with Magnific in San Francisco, featuring a walkthrough, a discussion on AI video workflows, and a live demo.
ChatGPT’s latest feature drop adds automatic quiz generation, natural-language restaurant reservation search, Google Drive access for paid users, and personalized suggestions.
MiniMax says Morgan Suo will discuss how the company’s H3 model is advancing AI video generation at ModCon in San Francisco on August 18.
Guizang (@op7418) has released guizang-sports-skill, a tool that lets Codex, Claude Code, and other agents analyze FIT or KML workout files and create route reports and shareable media.
A community-built desktop wrapper packages the official DeepSeek Harness experience into an install-and-run Electron app. The initial GitHub release supports Apple Silicon macOS and Windows x64, with local service management, plugins, and the Web UI retained.
Z.ai has announced GLM-5.3, a model focused on coding, agentic workflows, and cybersecurity. It is built through post-training on a 743-billion-parameter base model and is currently available via the GLM Coding Plan and ZCode, with API access and open weights planned in stages.