llm-anthropic 0.25

2026-04-16 · Source: Simon Willison's Weblog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

The `llm-anthropic` Python library, version 0.25, has been released, enhancing access to Anthropic's Claude series models. This update introduces support for the new `claude-opus-4.7` model, which features a `thinking_effort` setting of `xhigh`. Additionally, the release adds `thinking_display` and `thinking_adaptive` boolean options; `thinking_display` output is currently limited to JSON output or logs. The default `max_tokens` has been increased to the maximum allowed for each model, and the library no longer uses the deprecated `structured-outputs-2025-11-13` beta header for older models. This release was posted on April 16, 2026.

Key takeaway

For AI Engineers integrating Anthropic's Claude models, upgrading to `llm-anthropic 0.25` is crucial to access the new `claude-opus-4.7` model and its `xhigh` thinking effort. This update also provides new `thinking_display` and `thinking_adaptive` options, offering finer control and visibility into model processing, especially useful for debugging or optimizing complex prompts. Ensure your logging or output parsing supports JSON to fully utilize `thinking_display` summaries.

Key insights

The `llm-anthropic` library update adds new Claude models and enhanced control over model "thinking" processes.

Principles

• Model capabilities evolve with new releases.
• API headers can become obsolete over time.

In practice

• Utilize `claude-opus-4.7` for high-effort tasks.
• Explore `thinking_display` for debugging model thought processes.

Topics

• llm-anthropic
• Anthropic models
• Claude series
• claude-opus-4.7
• thinking_effort

Code references

• simonw/llm-anthropic

Best for: AI Engineer, Machine Learning Engineer, NLP Engineer

Related on AIssential

• llm-anthropic 0.25.1 — llm-anthropic 0.25.1 integrates Claude Opus 4.8, adds fast mode, and optimizes max_tokens defaults for better model interaction.
• Anthropic to launch Claude Opus 4.7 this week — Anthropic is expanding its AI offerings with a new model and a design tool, increasing competition in creative software.
• [AINews] Anthropic Claude Opus 4.7 - literally one step better than 4.6 in every dimension — Claude Opus 4.7 significantly advances AI capabilities in coding, vision, and complex task execution, despite potential token cost increases.
• Wtf Anthropic — Anthropic's compute miscalculation and opaque policies are alienating users, while OpenAI aggressively captures market share.
• Anthropic Unveils Updated Opus 4.7 Model — AI model development is rapidly advancing, with a clear distinction between general-purpose and specialized, high-risk applications.
• Live blog: Code w/ Claude 2026 — Anthropic is enhancing developer productivity and model autonomy through advanced agentic capabilities and platform integrations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Simon Willison's Weblog.