google / google/adk-docs

Found docs updates needed from ADK python release v2.9.0 to v2.9.1

Open Beginner friendly
#2,239 0 comments 0 reactions 0 assignees View on GitHub
docs updates
Dominant language
Shell
Stars
1.5k
Forks
1.3k
Avg merge
7d 1h
Merged PRs (30d)
34

Description

Compare URL: https://github.com/google/adk-python/compare/v2.9.0...v2.9.1

### 1. **Update Claude adaptive thinking documentation to mention summarized thoughts and version requirements.**

**Doc file**: docs/agents/models/agent-platform.md

**Current state**:
> Newer Claude models support *adaptive* extended thinking, where the model chooses
its reasoning depth itself rather than using a fixed token budget. On the native
Claude path, a negative `thinking_budget` maps to adaptive thinking.

**Proposed Change**:
> Newer Claude models support *adaptive* extended thinking, where the model chooses
its reasoning depth itself rather than using a fixed token budget, and returns its reasoning as summarized thoughts. On the native
Claude path, a negative `thinking_budget` maps to adaptive thinking. Adaptive thinking is required for Claude Opus 4.7 and later models, and recommended for Opus 4.6 and Sonnet 4.6 where standard "enabled" thinking is deprecated.

**Reasoning**:
The Python code for Anthropic models now adds `display="summarized"` to the adaptive thinking config. Additionally, the comments specify that adaptive thinking is now required for Claude Opus 4.7+ and recommended for Opus 4.6/Sonnet 4.6 where the old fixed budget "enabled" mode is deprecated. The documentation should be updated to reflect this behavior and provide this guidance.

**Reference**: src/google/adk/models/anthropic_llm.py

Contributor guide

Open the contributing guide

Research direction

Compare the v2.9.0 and v2.9.1 changes, then read docs/agents/models/agent-platform.md alongside src/google/adk/models/anthropic_llm.py. Update the adaptive-thinking description to cover summarized thoughts and the stated Claude version requirements. Done means the documentation matches the Python behavior and guidance described in the issue.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai, documentation
Issue type
Documentation
Difficulty
1/5
Estimated time
1-3 hours
Activity status
Active
Clarity
Clearly specified
Newbie friendliness
90/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.