Hugging Face switches language model inference to OpenRouter's smart routing
Hugging Face now routes language model requests through OpenRouter, which selects the best provider by live price, speed, and reliability and retries on failure.
How AI systems are built: agents, voice pipelines, integrations, and control points.
Hugging Face now routes language model requests through OpenRouter, which selects the best provider by live price, speed, and reliability and retries on failure.
Researchers used AlphaProof Code to mathematically prove Signal's code security properties, the first AI formal verification at implementation level. (136 characters)
Cloud Pathways is now GA on GKE, bringing Google DeepMind’s single-controller runtime to all customers with elastic training and multi-slice orchestration.
Foundry Local runs Qwen3.8-27B locally on NVIDIA GPUs with 24 GB VRAM, connects to GitHub Copilot, and works offline.
AWS blog introduces four-layer context engineering for production AI agents, with measured token savings of up to 97% in system prompts and 95% in code scripts.
Shopee built a specialized 245B MoE model Compass v3, now handling the majority of its AI traffic. It beats GPT-4o and other models on EcomEval benchmark. API tokens grew 113x in eight months according to the company.
Circle K deployed an AI agent on Amazon Bedrock to autonomously investigate DevOps alerts. Mean triage time fell from 45–60 minutes to under 15 minutes — a 67–75% reduction.
Cohere's Confidential Computing on Model Vault, with GPU memory protection and Oblivious HTTP encryption, is now available in public beta on Microsoft Azure with NVIDIA H100 GPUs.
Hardware-backed encryption for AI workloads in 36 zones with under 3.5% throughput loss.
Microsoft launches seven MAI models, previews Azure Cobalt 200 VMs, boosts Foundry IQ, introduces Scout and Discovery, and unveils Majorana 2.
The Gemini agent unifies AI tasks—coding, analysis, content creation—with dynamic sub-agents, but startups must contact sales for access.
ElevenLabs uses its own ElevenAgents for over 22,000 monthly support tickets, with 65% resolved end-to-end, 60-second response time, and 70% positive ratings – as self-reported.
Cohere reports that a strong lead model with cheap planning workers achieves near-top performance on coding tasks while cutting inference cost by 70%.
The full Run:ai platform and four modular components are now open source. An NGC Community Product offers installation without NVIDIA enterprise support.
Meta presented IcePack and 102.4-Tbps IceCube switches, alongside open protocols and networking architectures for large AI clusters.
Meta turns idle inference capacity in its fleet into better ads recommendations, using two elasticity levers that expand work without compromising baseline performance.
Local sandboxing for GitHub Copilot is now generally available for CLI, the Copilot app, and VS Code sessions using Agent Host. Policies restrict filesystem, network, and credential access.
NVIDIA says developers combine frontier AI models with Omniverse libraries to build simulation applications. The examples come from the company and have not been independently verified.
Google says several partners received access to two research prototypes during active emergencies. PDFM embeddings are commercially available in Preview. Figures are company-reported.
Google announced its Gemini agent at Gemini at Work 2026. The company says it plans tasks and returns finished work in existing apps. Pricing and availability were not stated.
Microsoft says GitHub Copilot will choose on-device or cloud models per task by end of October 2026. Local performance figures are vendor-reported.
GitHub's fine-tuned model reads surrounding code to flag likely credentials, including passwords without token formats. AI push protection is in private preview.
Microsoft Research Asia released Agent Lightning v1.0—a lightweight reinforcement learning framework for training AI agents using their existing harnesses without code reimplementation.
Canada Goose's AI agents autonomously resolved 89% of messaging inquiries and 15% of calls, improving service efficiency.
Salesforce introduced the Koa CRM model built on NVIDIA Nemotron 3 architecture and a new Enterprise AI Harness for AI governance and trusted operations.
Cohere released Embed 5, a new retrieval model series featuring a 128K-token context window and shared embedding space for enterprise use.
Developers can now build AI agents that navigate websites and interact with browser interfaces within an OpenAI-hosted environment.
NVIDIA has released Metropolis Blueprint VSS 3.3, featuring a new vss-build-vision-ai skill for creating video analytics agents from text prompts.
Google Cloud has released its Data Agent Kit, providing Model Context Protocol tools for integrating AI coding agents with cloud data services.
Meta has introduced new skills and connectors to its Muse AI agent to assist small business owners with operational automation.
OpenAI is rolling out Dots, persistent AI agents that manage third-party applications and perform proactive research for Pro and Enterprise users.
Claude autonomously discovered the array-associated reverse transcriptase (ART) system in bacteriophages by screening vast genetic datasets for repeating DNA patterns.
Google DeepMind has outlined plans for persistent, server-side memory in Private AI Compute, with encryption keys kept on users' devices.
Ringg has integrated GPT-5.6 to power its enterprise agent platform, reducing model costs by 90% while automating 65% of routine inquiries.
Amazon has added GPT-6 Sol and GPT-6 Luna to its Bedrock platform, offering lower API costs, prompt caching support, and improved factual accuracy.
Amazon has integrated Anthropic’s Claude Opus 5.5 into its Bedrock platform, featuring enhanced efficiency and new safety classifiers.
Load a GGUF model from the Hub by passing its ID and filename to from_pretrained. Packed inference is MPS-only for now and supports Qwen3.5 and compatible Qwen3.8 architectures.
Google uses its Mantis multi-agent framework to scan code pre-submit and autonomously patch vulnerabilities, preventing hundreds of issues monthly.
Google Cloud introduces Filestore agent volumes, a managed storage service designed for elastic and isolated AI agent workspaces in GKE.
NVIDIA reports that its latest Vera Rubin NVL72 hardware and software optimizations offer significant efficiency boosts for AI data centers.
Salesforce and AWS have introduced new capabilities for CRM data access, Slack-based AI agents, and foundation model support via Amazon Bedrock.
Salesforce and Google Cloud have expanded their partnership to integrate Salesforce's headless architecture with Gemini Enterprise and enable native interoperability.
Salesforce announced AIforce at Dreamforce, launching with Claudeforce, Slackforce and Agentforce Coworker. Salesforce in Claude is in beta. Pricing may change; availability varies by region and customer agreement.
Two Salesforce agents handle Siemens' inbound leads; the companies have announced a separate Teamcenter integration for sales and service.
OpenAI launched the Agents API public beta, featuring durable sessions, context compaction, and managed infrastructure for long-running AI agent execution.
OpenAI's GPT-Live-1 API allows developers to build full-duplex voice agents capable of simultaneous listening and speaking with native handling of interruptions.
The platform provides predictions for 9 billion single-nucleotide variants to assist researchers in interpreting human genome data.
Salesforce has introduced MCP server support to allow AI agents to securely access enterprise data and business logic without custom integrations.
Basis, Clay, and Exa Labs utilize AI agents for onboarding, account management, and developer tasks with mandatory human-in-the-loop review points.