Skip to content
HomeNewsConceptsGuidesToolbox
AboutSubscribeUA
Subscribe

AI Today Brief

The daily AI-engineering brief. Built in public. EN · UA.

XTelegramLinkedInYouTubeRSS

Follow AI Today Brief on LinkedIn for daily AI-engineering updates and the weekly “5 shifts that changed how developers work” PDF.

Explore

NewsDigestsConceptsGuides

Company

SubscribeAdvertiseAbout

Legal

Editorial policyAI disclosurePrivacyTerms

© 2026 AI Today Brief. All rights reserved.

  1. Home/
  2. News/
  3. Tools & releases/
  4. LLM Anthropic Plugin Adds Claude 5 Support and Server-Side Model Context Protocol Tools
Tools & releases

LLM Anthropic Plugin Adds Claude 5 Support and Server-Side Model Context Protocol Tools

August 5, 2026· 4 min read
OKCurated by Oleksandr Kuzmenko, AI Product Engineer·Updated August 5, 2026·Sources cited on every story
AI-assisted · editor-reviewed·How we use AI
LLM Anthropic Plugin Adds Claude 5 Support and Server-Side Model Context Protocol Tools

Simon Willison updated the llm-anthropic plugin, adding support for Claude 5 models alongside server-side tools like WebSearch, WebFetch, CodeExecution, and AnthropicMCP. The update integrates with llm>=0.32 to stream reasoning traces, tool execution events, and simplified thinking effort controls directly in the command line.

Impact: Medium

Why it matters

Integrating server-side MCP tools directly into the CLI removes the friction of building custom loops, streamlining command-line workflows for advanced reasoning models.

TL;DR

  • 01Adds support for Claude 5 models (fable-5, sonnet-5, opus-5).
  • 02Enables server-side tool calling via the -T flag.
  • 03Simplifies thinking effort configuration across five levels.

Key facts

Supported Models
claude-fable-5, claude-sonnet-5, claude-opus-5
Dependency Requirement
llm>=0.32
Server-Side Tools
WebSearch, WebFetch, CodeExecution, AnthropicMCP
Thinking Effort Options
low, medium, high, xhigh, max

Native Server-Side Tools and MCP Integration

The updated llm-anthropic plugin standardizes server-side tool invocations under the unified -T flag in the llm CLI. Developers can invoke WebSearch, WebFetch, CodeExecution, and AnthropicMCP directly without maintaining custom client-side execution loops. Server-side tool execution and reasoning events stream as typed events when paired with llm>=0.32.

Simplified Reasoning and Thinking Controls

Extended thinking configuration replaces options like thinking_budget, thinking_display, and thinking_adaptive with streamlined thinking and thinking_effort settings (low, medium, high, xhigh, max). For claude-sonnet-5 and claude-opus-5, thinking can be toggled off via -o thinking 0, while claude-fable-5 keeps thinking enabled permanently. Use -R or --hide-reasoning to suppress reasoning tokens from standard output and logging streams.

Try it in 2 minutes

llm -m claude-sonnet-5 -T WebSearch "Latest AI developments"

bash

✓ When to use

  • When running Claude 5 models and requiring server-side tool use via the LLM command line tool.

✕ When NOT to use

  • When you are using older versions of the LLM CLI below 0.32 without updating.

What to do today

  • →Upgrade your llm package to version 0.32 or newer.
  • →Use the -T flag to trigger server-side tools like WebSearch or CodeExecution.
#llm#llm-anthropic#WebSearch#WebFetch#CodeExecution#AnthropicMCP

Sources

  • LLM access to models by Anthropic, including the Claude series
ShareShare on XShare on LinkedIn
← Previous storyRust Monorepo Formulates Policy for AI Code Contributions and PR ReviewsNext story →Inspecting Codex Request Payloads Reveals 9,400 Token Overhead per Prompt

Related stories

  • Tools & releasesShieldFont Defends Web Content Against AI Scraping via Font Glyph Obfuscation
  • Tools & releasesJFrog Exposes Batch of Fabricated SQLite CVEs Generated by LLMs
  • Tools & releasesAurora: Open-Source Self-Hosted AI Gateway in Go for Multi-Provider Routing
  • Tools & releasesCursor Removes Dollar Cost Data from Usage Dashboard and CSV Exports

Email digest

Get the morning AI brief

One email a day — the stories that matter for engineers, founders and tech leads. Human-edited, with links to primary sources.

  • ✓120+ sources scanned daily
  • ✓Edited by a human
  • ✓1 email per day
  • ✓EN + UA

By subscribing you agree to the privacy policy.