Roblox、ゲーム開発にAIエージェントを接続させる「Studio MCP Server」のアップデートを発表、
moguravr.comRoblox announced an update to Studio MCP Server, enabling AI agents connected via the MCP protocol for game development workflows.
DProの広告データ分析から画像・動画制作まで、Codex／Claude Codeなどでつなぐ「Pro ai MCP」を提供開始
asahi.comKASHIKA launched Pro ai MCP, an AI tooling integration service for Codex/Claude Code environments that enables image and video generation from advertising data analysis.
大语言模型推理框架怎么选？主流推理开发框架选型指南（2026）
ithome.comChinese tech article discussing how to choose large language model inference frameworks in 2026, with llama.cpp mentioned as a mainstream option for local LLM deployment and inference optimization. Published 2026-08-12.
Civic Marketplace Connectors Bring Local Government Procurement Into Claude, ChatGPT and Copilot
finance.yahoo.comCivic Marketplace is integrating its public sector procurement capabilities directly into AI assistants including Copilot, enabling local government workers to access procurement tools through their preferred LLM interface.
Mistral and Humain Collaborate to Advance Sovereign AI
developingtelecoms.comMistral AI and Humain announced a strategic collaboration focused on advancing sovereign AI capabilities, leveraging Mistral's models across their product portfolios.
Mistral’s Le Chat chatbot gets a productivity push with new ‘deep research’ mode
finance.yahoo.comFrench AI lab Mistral introduced new features to its Le Chat chatbot, including a 'deep research' mode that enhances productivity and information gathering capabilities.
Ottawa teams up with startup Cohere to bring AI to the public sector
finance.yahoo.comOttawa partners with Cohere startup to deploy sovereign AI solutions for the public sector.
NVIDIA launches open model Nemotron 3.5 Lightning; working on bigger Nemotron 4: report
msn.comNVIDIA expands its Nemotron 3 open model family with Nemotron 3.5 Lightning, the highest-efficiency model for long-running agentic AI workloads, while working on trillion-parameter Nemotron 4.
Gemma hit 1 billion downloads without a server — NASA already took it to orbit
martincid.comGoogle DeepMind's Gemma family of open AI models has crossed one billion cumulative downloads, with outside developers using the models for various applications including NASA satellite image analysis.
Liquid AI与Hugging Face联手，LFM2.5 DSpark草稿模型推理性能大幅提升
msn.cnLiquid AI and Hugging Face collaborate on LFM2.5 DSpark, significantly improving inference performance of the draft model.
Qwen, da Alibaba, lança modelo de IA Qwen3.8-Flash com custos de treinamento menores
msn.comAlibaba's AI division launches new Qwen3.8-Flash model with reduced training costs, offering competitive performance against recent models from Anthropic and DeepSeek.
微信竟把“隐形AI”开源了，不过还是基于Qwen训练！
news.qq.comWeChat's visual team has open-sourced models based on Qwen3.5, with 2B, 4B, and 9B versions supporting text, images, video, visual documents, and interleaved multimodal inputs.
商汤大装置为智谱GLM-5.3-Flash提供国产算力与Token服务
yicai.comSenseTime's large-scale computing platform provides domestic compute power and token services for Zhipu AI's newly launched GLM-5.3-Flash model, marking another milestone in the commercialization of domestic computing infrastructure.
智谱开源GLM-5.3-Flash多模态模型：性能超越前代，国产芯片高效支撑
sohu.comZhipu AI announces the official launch and open-source release of GLM-5.3-Flash, a 32B-A18B parameter multimodal model that outperforms previous generations in benchmark scores while running efficiently on domestic Chinese chips.
智谱确认Ox Alpha为新一代GLM模型，使用量已达DeepSeek两倍
sohu.comZhipu AI confirms Ox Alpha as the new generation of its GLM series model, which has attracted significant attention in the AI model community. The weights are being made publicly available and remain free for users.
ChatGPT may be changing how you're writing
tech.yahoo.comSFGate article about ChatGPT's impact on writing styles and AI detection challenges.
Google launches Gemini Omni 1.1 flash with major video creation upgrades
thenews.com.pkGoogle has introduced Gemini Omni Flash, a significant move aimed at launching an upgraded generative AI model with major video creation upgrades.
Claude Mythos Hype vs. Reality: What Security Teams Need to Know
darkreading.comDiscussion of Anthropic's Claude Mythos rollout and what security teams need to know about the hype versus reality of this AI tool.
Nvidia says Groq 3 LPX now in full-scale production
seekingalpha.comNvidia says Groq 3 LPX is in full production, boosting AI inference to 3,400 tokens/sec and faster agentic AI capabilities.
OpenAI's rogue AI model incident was worse than we thought
theverge.comTwo new reports offer nearly 130 pages of details on the OpenAI-Hugging Face cybersecurity incident, many previously undisclosed aspects.