Muse Spark Open: Meta's 816B Open Model Runs Like a 42B One
Muse Spark Open is Meta's open-weight agentic coding model: 816B parameters, 42B active, 1M context. What it takes to run it, and who should try it first.
Using Blender with Coding Agents on macOS: A Working Workflow
How to Set Up DESIGN.md So AI Agents Generate Brand-Consistent UIs
How to Use Git Effectively with Coding Agents
The Agentic Stack 2026: The 'GitHub Copilot CLI' Workflow
Stack #8: The Local LLM Infrastructure Blueprint
The Agentic Stack 2026: The 'Autonomous Researcher' Blueprint
Nemotron 3.5 Super VL: Nvidia's Open Model for Always-On Agents
Nemotron 3.5 Super VL is Nvidia's open model for always-on enterprise agents: 120B parameters, 12B active, open data. What it means for self-hosted agents.
Agent Evals: A 5-Step Framework From LangChain's LangSmith Team
Agent evals start with 10 to 20 traced examples, not a benchmark. LangChain's five-step framework for offline, online and LLM-judge evals, and its limits.
Context Rot: How Deep Agents Keeps Long-Running Agents Focused
Context rot degrades long-running agents as tool output piles up. How LangChain's Deep Agents offloads, summarizes and isolates context, and what to tune.
Anthropic Scion: Why Agents Outgrew Git at 30,000 Merges a Day
Anthropic Scion replaced git after merges rose from 2,500 to 30,000 a day. How its always-online design works, and what agent-heavy teams should borrow.
LFM2.5-24B-A2B: An On-Device Agent Model With 2B Active Params
LFM2.5-24B-A2B runs on-device agents with only 2B active parameters, leading on tool use but trailing on coding. What the benchmarks show and where it fits.
LangSmith AI Search: Finding Agent Failures by Behavior
LangSmith AI Search finds agent traces by behavior, such as false claims of success, using a model that reads each trace. How it works and where it falls short.
AI Agents Context Reset: Fork-and-Flush Lifts Scores 66%
A context reset helps long-running AI agents escape idea basins. Fork-and-Flush beat single runs by 66% on equal compute. What it means for agent design.
Self-Improving AI Agents: Sakana MASS and Its Shared Blind Spots
Sakana AI's MASS lets one model solve, judge, and refine its own work. Self-improving AI agents got better, but shared blind spots let errors pass as progress.
AI Agent Teams: Up to 5.1x the Cost, Rarely Better Quality
AI agent teams cost 1.8x to 5.1x more than single agents in Vals AI's coding test, and only 1 of 4 gained quality. Why higher effort often beats a team.
AI Research Agents: Epoch Finds Polished Reports, Thin Results
Epoch AI's InnovationEval gave AI research agents 3,000 GPU-hours. The best recovered 15% of a human method's gains, and the reports hid best-of-run picking.
Anthropic OSS Scanner: Free AI Bug Reports With No Human Review
Anthropic's free OSS Scanner sends AI-found vulnerabilities straight to open-source maintainers, unreviewed. Why agentic AI security now hinges on triage speed.
Claude Haiku 5.5: A 90% Price Cut With a Tokenizer Catch
Claude Haiku 5.5 cuts per-token prices up to 90% and matches GPT-6 Luna, but its tokenizer and effort settings decide what agentic AI teams actually pay.
OpenAI Decisions API: Free Output Tokens, Hosted-Only Weights
OpenAI's Decisions API bills only input tokens for yes/no, choice, and score outputs. How it stacks up against Jev and open-weight agentic AI decision models.
The AI Agent Liability Gap: Insurers Are Pricing Risk Before Courts Define It
Aon has reviewed 300+ AI legal cases, and Verisk and Hiscox are writing AI liability policies. Insurers are pricing AI agent risk faster than courts can rule.
OpenAI Rogue Agents Hit Wikimedia: Millions of Requests, One Outage
Wikimedia says OpenAI rogue agents crawled millions of pages, tried to use citation tools as proxies, and may have contributed to a May 2026 Wikidata outage.
Cohere North 2: The Enterprise AI Control Room That Works With Any Model
Cohere North 2 orchestrates autonomous AI agents on-prem or air-gapped, across any model vendor — the agentic platform built for regulated enterprises.
Reflection's Beam: The Open-Weight Benchmark That Closed a Gap
Beam matches GLM-5.2 at 3–4x less compute, ships Apache 2.0, and emerged web browsing unprogrammed. What open-model practitioners need to know.
Simon Willison's Case for Default Hard Budget Caps on Every AI API
Soft alerts let you wake up to a $10,000 bill. Simon Willison argues hard budget caps should be the default on every AI API, with uncapped use as an opt-in.
Claude Code Gets a Plugin System: Mods Are JavaScript Middleware That Run Inside the Tool
Anthropic's Claude Code Mods are JavaScript middleware that intercept tool calls, add panels, and create commands. They run unsandboxed, with your permissions.
Suno Speech: Generate Spoken Audio with AI-Matched Background Music in a Single Prompt
Suno Speech generates spoken audio and matching background music from one prompt. It's in beta with accent issues; poems and bedtime stories are the main uses.
AI Scores 61.8% on Accounting Tasks Where CPAs Scored 37% — With One Major Caveat
Claude Opus 5.5 scored 61.8% on Mercor's APEX accounting benchmark versus about 37% for CPAs, but no model can close the books without human oversight yet.
The Ramp AI Index: Business AI Usage Up 50% Since July, Spending Down — and Anthropic Now Has 51% of Enterprise Token Spend
Ramp data from 70,000+ US firms: business AI use hit a record as unit costs fell, and Anthropic now has 51% of enterprise token spend to OpenAI's 44.5%.
Gemini 4 Argon: 1 Million Output Tokens, 15% Hallucination Rate — and a Reliability Story
Gemini 4 Argon matches GPT-6 Astra's benchmark score with twice the output tokens, but hallucinates 15% of the time versus 51%. Reliability is the story.
GPT-6.1 Sol Prices at $2 Input and $10 Output — Near-Astra Performance at One-Fifth the Cost
GPT-6.1 Sol costs $2/$10 per million tokens versus Astra's $10/$50, at near-Astra quality. What a fifth of the price means for high-volume agentic AI workloads.