claude-code - 💡(How to fix) Fix Feature Request: Enable thinking.display for Claude Code subscription sessions [1 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
anthropics/claude-code#52376Fetched 2026-04-24 06:08:50
View on GitHub
Comments
1
Participants
2
Timeline
3
Reactions
0
Timeline (top)
labeled ×2commented ×1

This is not about exposing raw chain-of-thought. The "summarized" display mode already exists and produces concise reasoning summaries appropriate for developer consumption. The infrastructure is built — it just needs to be enabled for subscription auth.

Root Cause

This is not about exposing raw chain-of-thought. The "summarized" display mode already exists and produces concise reasoning summaries appropriate for developer consumption. The infrastructure is built — it just needs to be enabled for subscription auth.

RAW_BUFFERClick to expand / collapse

API users can set thinking.display: "summarized" to receive summarized reasoning content from Claude models with adaptive thinking. Claude Code has a showThinkingSummaries setting (visible in the settings schema) but the server does not honor the thinking.display parameter for subscription-authenticated sessions — only API key sessions receive thinking content.

Use case

As AI-generated code enters production systems, developers need to understand why the model made specific decisions — not just what it output. Debugging AI-authored code without access to the reasoning that produced it is like debugging a binary without source. This is especially relevant as agent workflows scale: Uber reported 11% of their live backend code is now AI-generated, and that number is climbing.

Summarized thinking content would enable:

  • Post-session audit of model reasoning for critical code changes
  • Debugging when AI-generated code produces unexpected behavior
  • Compliance and explainability in regulated environments where automated decisions require documentation
  • Developer learning by understanding how the model approaches complex problems

Current behavior

  • CC sends redact-thinking-2026-02-12 as an anthropic-beta header by default
  • When showThinkingSummaries is enabled in settings, CC removes the redact flag and sets thinking.display: "summarized" on the request
  • The server returns a thinking block with an empty thinking field and a signature (used for token accounting)
  • The same request structure returns summarized thinking text for API key sessions

Requested behavior

Honor thinking.display: "summarized" for CC subscription sessions the same way it is honored for API key sessions. The showThinkingSummaries setting already exists in CC — the server just needs to return the content.

Context

This is not about exposing raw chain-of-thought. The "summarized" display mode already exists and produces concise reasoning summaries appropriate for developer consumption. The infrastructure is built — it just needs to be enabled for subscription auth.

extent analysis

TL;DR

The server needs to be updated to honor the thinking.display: "summarized" parameter for subscription-authenticated sessions, similar to how it is handled for API key sessions.

Guidance

  • Verify that the showThinkingSummaries setting is enabled in the Claude Code settings schema for the subscription session.
  • Check the server-side logic to ensure it correctly handles the thinking.display: "summarized" parameter for subscription-authenticated sessions, potentially by removing the redact-thinking-2026-02-12 header.
  • Compare the request structures for API key sessions and subscription sessions to identify any differences that might be causing the discrepancy in behavior.
  • Test the server's response with the thinking.display: "summarized" parameter set for both API key and subscription-authenticated sessions to confirm the issue is resolved.

Notes

The fix may require updates to the server-side logic to handle subscription-authenticated sessions consistently with API key sessions. The existing infrastructure for summarized thinking content and the showThinkingSummaries setting suggests that the necessary components are already in place.

Recommendation

Apply a workaround by modifying the server-side logic to honor the thinking.display: "summarized" parameter for subscription-authenticated sessions, as this will enable the desired functionality without requiring significant changes to the existing infrastructure.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING