The latest from the AI and MCP ecosystem, curated daily.
A practical engineering account of building a self-healing deployment pipeline for LangChain's GTM Agent — after every deploy, it detects regressions, triages whether the new code caused them, and automatically opens a PR with a fix. No manual intervention until review time. A concrete example of agents managing agents in production CI/CD workflows.
LangChain's evals show open models like GLM-5 and MiniMax M2.7 now match closed frontier models on core agent tasks — file operations, tool use, and instruction following — at a fraction of the cost and latency. The post shares the benchmark results and explains how to start using these models in Deep Agents. A significant data point for teams weighing open vs. closed models for production agents.
OpenAI added pay-as-you-go pricing for Codex on ChatGPT Business and Enterprise plans, removing the need for a fixed seat commitment to get started. Teams can now scale Codex usage up or down without upfront contracts. A meaningful change for teams that want to pilot Codex without a large spend commitment.
Google introduced new controls in the Gemini API for balancing cost and reliability — including tiered pricing options, fallback routing between model variants, and latency budgets for production deployments. Addresses the practical production concerns teams face when scaling Gemini-based applications beyond prototypes.
Google released Gemma 4 — their most capable open model family yet, with multimodal understanding and strong performance across reasoning, coding, and instruction following tasks. Designed to run on-device and at edge scale, Gemma 4 closes the gap with frontier closed models while remaining fully open-weight. A significant update to the most widely-used Google open model series.
Cursor launched Cursor 3 — a unified workspace for building software with agents, bringing together the editor, cloud agents, and background automations into a single coherent product. The release represents a significant architectural shift from IDE-with-AI-features toward an agent-first development environment. The biggest Cursor product update to date.
Google released Gemma 4 — their latest open-weight model family with frontier multimodal capabilities designed to run on-device. Gemma 4 brings vision understanding, multilingual support, and significantly improved reasoning to a form factor that fits on consumer hardware. A major open-weight model release that expands what's possible for local AI development.
Anthropic outlines three patterns for building applications that keep pace with Claude’s evolving capabilities: lean on what Claude already knows, continuously question which harness logic Claude can now handle itself, and set boundaries carefully. The post highlights how assumptions baked into agent harnesses go stale as models improve, and why giving Claude tools like bash and a code executor lets it orchestrate its own actions more efficiently. Essential reading for anyone building production agents.
H company released Holo3 — a new computer use model that claims state-of-the-art performance on browser and desktop automation benchmarks. The post covers architecture decisions and benchmark results showing Holo3 outperforming existing computer use models on key tasks. A significant new entrant in the fast-moving computer use agent space.
Google launched an official MCP server for Gemini API documentation, letting coding agents access up-to-date API reference directly during development without hallucinating outdated method signatures. Plug it into any MCP-compatible coding agent (Claude Code, Cursor, Codex, etc.) to ground API calls in real documentation. A practical quality-of-life improvement for developers building with Gemini.
OpenAI announced a $122 billion funding round to expand frontier AI globally, invest in next-generation compute infrastructure, and meet growing demand for ChatGPT, Codex, and enterprise AI products. One of the largest private funding rounds in history, signalling continued massive investment in AI at the frontier.
Hugging Face released TRL v1.0 — a major milestone for their post-training library covering RLHF, DPO, PPO, and other alignment techniques. The v1.0 release signals API stability and production readiness for teams fine-tuning and aligning open-source models. The go-to library for post-training just became officially stable.
Anthropic published a Compliance API that surfaces platform activity for auditing and governance. For engineering teams this simplifies incident analysis and compliance reporting by providing structured access to usage and action logs, reducing manual forensics. It matters for teams building agentic systems where auditability and traceability are required for safety and regulatory reasons.
A comprehensive practical checklist for agent evaluation covering error analysis, dataset construction, grader design, offline and online evals, and production readiness criteria. One of the most complete guides to building a rigorous agent eval pipeline from scratch. Useful for any team that needs to move from "our agent seems to work" to confident production deployment.
Cursor published an arXiv technical report on training Composer 2 — covering continued pretraining on Kimi K2.5, large-scale RL in realistic Cursor sessions, and CursorBench results alongside public benchmarks. The report gives the research community a detailed look at how a production AI coding model is actually trained and evaluated.
Hugging Face published a guide on running OpenClaw agents locally with open-weight models — covering model selection, setup, and the tradeoffs between hosted and self-hosted agent deployments. Relevant for developers wanting to reduce API costs or run agents in air-gapped environments using open-source models.
LangChain shares their internal methodology for building evals for Deep Agents — how they source data, design metrics, and run targeted experiments over time to make agents more accurate and reliable. The key principle: the best evals directly measure a specific agent behaviour you care about, not proxies. Practical guidance backed by LangChain's own production experience.
LangChain introduces Agent Middleware — a pattern for customising agent harnesses by injecting logic between the LLM and its environment without modifying core harness code. Covers the design pattern, use cases (logging, auth, rate limiting, retry logic), and how it enables teams to build application-specific harnesses on top of LangChain's Deep Agents framework.
Google released Gemini 3.1 Flash Live — a low-latency model variant optimised for real-time conversational agent use cases, with sub-500ms response times for voice and interactive chat. Covers the architecture, streaming API, and how to build agents that feel instant rather than waiting on generation. Key for anyone building voice or real-time agent interfaces.
Cursor applies online reinforcement learning to Composer — serving new model checkpoints to production and using real user interactions as reward signals to ship improved checkpoints multiple times per day. A practical account of how they close the loop between production usage and model training at high cadence. Fascinating engineering for anyone building RL-trained coding models.
