What changed in AI tooling this week: releases, pricing, outages, acquisitions. Every item links to its source.
The RL trained model treats a trace like a repo, grepping spans to catch quiet errors such as an agent marking work done while tests still fail, and on Laminar's own 523 trace benchmark it edges GPT…
The pairing scores 68.1 on Artificial Analysis's Coding Agent Index v1.5 (DeepSWE 72.6, Terminal Bench 4 62.1, SWE Atlas QnA 69.6) at $9.01 and about 55 minutes per task, effectively tied with Muse C…
Per The Information, Anthropic is testing a model that infers your real goal from thin instructions (one tester says it traces a single bug to the broader problem and offers a fuller fix), Google is…
Foundry agents now get a managed cloud browser (via Playwright Workspaces, provisioned as a separate Azure resource) that can reach public sites and private VNet apps, exposed to prompt or hosted age…
On OpenAI's new MultiLangBench, where a model builds a SQLite compatible engine in 20 languages, GPT 6 Astra averages an 87.4% test pass rate (COBOL jumps from 10.9% to 72.1%) versus 26.8% for GPT 5.…
Built on MCP, the connectors let an agent carry engineering work from CAD (CATIA, NX) through meshing (ANSA, HyperMesh) to simulation (Abaqus) across Windows workstations and Linux clusters, with eng…
The filesystem, memory, sequential thinking, and everything servers move to TypeScript SDK v2 and the 2026 07 28 spec while still answering older clients, and the Python fetch, git, and time servers…
Workspace chats can edit the dev checkout, run commands, and read service logs (asking before Git pushes by default, with optional approvals for shell, web, and platform changes), while organization…
MiniMax says the agent and coding model, which ran anonymously as "Space Bunny" and served over 200 trillion tokens, is nearly 3x faster than its predecessor (170+ tokens/sec median) and cuts agentic…
Mercury 3 generates text by parallel refinement instead of token by token, and Inception says it matches or beats GPT 6 Luna and Claude Haiku 5.5 while running 4–8x faster. It's OpenAI API compatible…
Agents can subscribe to six events at launch (new users, failed function executions, finished function and site deployments, new rows, and uploaded files) and get them over an open stream locally or…
Projects like dumbtv and cleanyourtv let you aim a coding agent at an Android TV, Google TV, or Fire TV over Wi Fi to remove sponsored ads, disable ACR viewing tracking, and swap in a clean launcher,…
Amnon Shashua's doubleAI calls Maxwell a "code execution model": instead of writing source like Claude Code, it watches a live program on CPU or GPU, emits faster binaries for the hardware and data i…
Subject to stockholder approval, a wholly owned LLC gets 5% of Anthropic's shares now, plus another 2% at the board's discretion for every 50% rise in value above a baseline, and Anthropic says none…
Full BYOC (private preview) puts compute, storage, and the Agent harness in your cloud, BYOC for Published Apps (early access) keeps the publish button but hosts apps in your tenant with private endp…
In a Work or Codex chat, @Chrome (or Edge, Brave, Opera, Vivaldi) lets ChatGPT fill forms, send messages, download files, and group tabs on sites where you're already logged in, stopping before submi…
You describe a task like ticket routing or text classification, a large model synthesizes the labels and handlers at build time, and a small local model then answers at about 60 ms per record with no…
Pillar Security found that OpenCode 1.0.137–1.18.32 on Windows/WSL passed a remote MCP server's authorization endpoint straight into PowerShell via the open package, so a $(...) HTTPS URL could run c…
New validate pipeline checks YAML against Buildkite's schema without starting a build, and compare builds reports newly failing, recovered, and still broken steps with log tails for up to three failu…
An internal Anthropic model worked out in under half an hour that the NSA supplied seeds for P 192 through P 521 came from the C lrand48 / drand48 generator seeded with Unix timestamps from 1997 to 1…