Fetching from the wire…
Public story · 2026-08-07 · high
One route covers seven AWS regions with a plain model ID; the other needs a London-only ARN for the newest Claude models.
Why now: AWS itself calls the seven-versus-one region split decisive for anyone configuring single-region Claude Code, per its blog post.
AWS published two ways to keep Claude Code inference inside one region instead of one geography, according to a post from its machine learning team. Cross-Region inference is Bedrock's default because it routes prompts to whichever region has capacity, and that default breaks residency rules for customers who need every request to stay put, not just stay in-country.
The first path sets the environment variable CLAUDE_CODE_USE_MANTLE=1 alongside AWS_REGION and pins the model by its plain ID. AWS supports it in seven regions: Ireland, Stockholm, Tokyo, Melbourne, N. Virginia, Ohio and Oregon.
The second path uses classic Bedrock with an application inference profile ARN. Only eu-west-2, London, has that profile wired up for Opus 4.6 and Sonnet 4.6, per AWS's post.
AWS calls that split decisive for most people setting this up. Outside London, anyone who wants Opus 4.6 or Sonnet 4.6 pinned to a single region has to take the Mantle path, not the ARN one, because the ARN path isn't wired up past eu-west-2. That's a bet worth tracking: if AWS adds inference profiles for the newest models in more regions, the seven-versus-one gap closes and the ARN path stops being a London-only workaround. Until then, it is.
Each link below shares sources, entities, or timing with this story.
A GitHub repo cataloging Claude Code tips doesn't normally warrant a top story. But shanraisshan/claude-code-best-practice at 53.4K stars isn't a tips list anymore. It's the de facto reference for how an entire generation of developers is learning to work with AI coding agents...
Anthropic launched Claude Sonnet 4.6 claiming performance comparable to Opus 4.5 at $3/$15 per million tokens (vs. Opus's $5/$25). SWE-bench Verified: 79.6% (near Opus 4.6's 80.8%). OSWorld-Verified: 72.5% (tied with Opus 4.6's 72.7%). 1M token context window in beta. Now the...
If the GitHub story is the supply side of agentic coding straining, this is the demand side. Anthropic opened its Seoul office, its third in Asia-Pacific after Tokyo and Bengaluru, and the launch came bundled with enterprise deployments that should end the "is this still a pil...
AWS's August 21 post stages agent tool governance as Connect, Control, Catalog and Harden, from one SSO-backed MCP endpoint for a 1-20 user pilot through identity-aware authorization with PII redaction and self-service tool publishing at 100+ users. It supports Cognito-backed...
Buried in the 2.1.233 notes. The harness is removing explicit planning scaffolding for the newer tier while keeping it for older models, betting these models track multi-step work internally and that a forced todo list burns tokens without improving completion. Claude Code Rel...
AI Now Institute researchers Boyan Milanov and Heidy Khlaaf demonstrated turning a coding agent doing vulnerability review into the execution vector, planting hidden binaries disguised as build artifacts alongside a deceptive README.md. The payload worked unchanged on Sonnet 5...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.