ATRIUMsearch → argument graph
Article · 2026-08-04 · 4 moments

llm-anthropic 0.26

Release: llm-anthropic 0.26 Includes new features enabled by LLM 0.32: New models: claude-fable-5, claude-sonnet-5, and claude-opus-5. #75, #76 Added server-side tools for WebSearch, WebFetch, CodeExecution, and AnthropicMCP, available through LLM's -T interface or Python tools=. The previous -o web_search* options have been removed in favor of -T WebSearch. #79 Upgraded to llm>=0.32. Reasoning, tool calls, tool results, and server-side tool results now stream as typed events. Reason ✦ AI generated

Read original ↗
GUEST llm-anthropic maintainerHOST Simon Willison
01
Claim◆

Extended thinking is simplified to two options—thinking and thinking_effort (low, medium, high, xhigh, or max)—with Claude 5 models thinking by default, while Fable 5 always thinks regardless of configuration.

Thinking configuration is simplified, with Claude 5 models defaulting to thinking and Fable 5 always thinking.

transcript

llm-anthropic maintainer: Simplified extended thinking to thinking and thinking_effort (low, medium, high, xhigh, or max). Claude 5 models think by default; -o thinking 0 disables thinking for Sonnet 5 and Opus 5, while Fable 5 always thinks. -R/--hide-reasoning now omits reasoning from responses and logs. The thinking_budget, thinking_display, and thinking_adaptive options have been removed. #80

provides context · 1

02
Claim

Server-side tools for WebSearch, WebFetch, CodeExecution, and AnthropicMCP are now available through LLM's -T interface or Python tools=, replacing the previous -o web_search* options.

New server-side tools replace the older web_search options, accessible via the -T interface.

transcript

llm-anthropic maintainer: Added server-side tools for WebSearch, WebFetch, CodeExecution, and AnthropicMCP, available through LLM's -T interface or Python tools=. The previous -o web_search* options have been removed in favor of -T WebSearch. #79

extends · 2provides context · 1

04
Fact

With llm>=0.32, reasoning, tool calls, tool results, and server-side tool results now stream as typed events, and reasoning displays to standard error by default.

The upgrade to llm 0.32 enables typed event streaming for reasoning and tool interactions.

transcript

llm-anthropic maintainer: Upgraded to llm>=0.32. Reasoning, tool calls, tool results, and server-side tool results now stream as typed events. Reasoning for llm CLI prompts now displays to standard error unless you pass --hide-reasoning/-R.

extends · 1provides context · 1

Highlight slides
Three New Claude 5 Models Released✦ from: Three new Claude 5 models—claude-fable-5, claude-sonnet-5, and claude-opus-5—are now available in llm-anthropic 0.26.Thinking Configuration Simplified✦ from: Extended thinking is simplified to two options—thinking and thinking_effort (low, medium, high, xhigh, or max)—with Claude 5 models thinking by default, while Fable 5 always thinks regardless of configuration.Model Behavior Varies by Tier✦ from: Extended thinking is simplified to two options—thinking and thinking_effort (low, medium, high, xhigh, or max)—with Claude 5 models thinking by default, while Fable 5 always thinks regardless of configuration.New Reasoning Display Control✦ from: Extended thinking is simplified to two options—thinking and thinking_effort (low, medium, high, xhigh, or max)—with Claude 5 models thinking by default, while Fable 5 always thinks regardless of configuration.
Related episodes
Article · ByteByteGo Newsletter

MCP vs A2A vs ACP: How AI Agents Actually Talk to Each Other

→
Video · Acquired

Disney: The Renaissance and the Empire

→
Video · The Cognitive Revolution

Alignment with Awakening: Davidad on Moral Realism, AI Wisdom, & why His p(Doom) is Down to 5%

→