<p><strong>Claude Sonnet 4 and Opus 4 retire June 15</strong> — production apps calling the <code>-20250514</code> model IDs will break, so migrate to Sonnet 4.6 or Opus 4.8 now before the deadline. Microsoft Work IQ APIs go GA June 16 with audited access to Microsoft 365 org intelligence for agents, and LangGraph Platform gains n...
<p><strong>Gemini 3.5 Pro</strong> is closing in on general availability with a 2M-token context window and ~$15/$60 per MTok pricing, setting up a new orchestrator-vs-sub-agent cost calculus for multi-agent builders. Meanwhile, <strong>ServiceNow</strong> expanded AI Control Tower across 30 enterprise integrations to govern agents anywhere they run, and ...
<p>Google's <strong>Gemma 4 12B</strong> lands as a free, encoder-free multimodal model (text, image, audio, video) on 16 GB VRAM under Apache 2.0, while Cohere's <strong>Command A+</strong> becomes the first fully open enterprise-grade large model — giving builders two credible self-hosted paths in one week. Meanwhile, OpenAI is sunsetting Agent Builder ...
<p><strong>OpenAI's Codex app</strong> now controls Windows desktops — agents can see, click, and type in any live application, raising the bar for fleet sandboxing and security policy. Meanwhile, <strong>Amazon Bedrock AgentCore</strong> landed its Python SDK on PyPI with persistent S3/EFS file mounts, and the <strong>MCP spec released a stateless HTTP r...
<p><strong>Anthropic filed a confidential S-1 with the SEC</strong> at a $965 billion valuation and $47 billion revenue run rate, racing OpenAI to public markets. Meanwhile, <strong>Microsoft's MAI-Code-1-Flash</strong> rolled out to all GitHub Copilot plans — its first in-house coding model built without OpenAI data — as Copilot switches to token-based A...
<p><strong>Microsoft Build 2026</strong> ships <strong>Windows Agent Framework 1.0</strong> (MIT) and <strong>Azure Agent Mesh</strong> for federated multi-agent orchestration, while <strong>GPT-5.5</strong> arrives at roughly half the cost of comparable frontier models. Operators also face two urgent deadlines: <strong>GPT-4.5 retires June 27</strong> an...
<p><strong>Claude Opus 4.8</strong> lands with 1M-token context and native dynamic multi-agent workflows, while Google replaces Vertex AI with the unified <strong>Gemini Enterprise Agent Platform</strong> (ADK v1.0, A2A v1.0 in production). Meanwhile, the EU AI Act transparency deadline hits August 2 and Colorado's AI law takes effect June 30 — two compli...
<p>Cohere's <strong>Command A+</strong> lands as the first Apache 2.0 frontier-class model — 218B sparse MoE running on two H100s with built-in citation spans for RAG pipelines. Meanwhile, <strong>ServiceNow Action Fabric</strong> goes GA, letting any AI agent execute governed enterprise workflows through an MCP server with a full audit trail.</p>
<p>Amazon Bedrock AgentCore now lets agents <strong>autonomously pay for APIs and MCP services</strong> via Coinbase and Stripe wallets — the first native payment primitive from a major cloud. Meanwhile, <strong>AG2 pushes toward a stable v1.0</strong> with an event-driven core, SAP backs n8n at a $5.2B valuation to embed workflow automation inside its en...
<p>Anthropic launched <strong>MCP tunnels</strong> and <strong>self-hosted agent sandboxes</strong> (public beta) to let agents reach private infrastructure without inbound firewall exposure — with Cloudflare, Modal, and Vercel as day-one sandbox partners. Meanwhile, <strong>SGLang's RadixAttention</strong> hits 400,000-GPU production scale at 6.4× vLLM t...