Follow Discord
Sweep 22 Oct 2026 · 21:04Z Build v2.1.309 528 read Stable v2.1.308 Latest v2.1.309 Next v2.1.309 Feeds RSS JSON llms.txt llms-full.txt Unofficial
HomeDocsClaude Developer Platform 23 Oct 2026
One change · api

preserved-thinking changedbuild-with-claude/preserved-thinking

Nearest release: v2.1.309, published 3 hours before this site recorded the change. Shown because the two are within 24 hours of each other. Nothing here says the release caused the edit.

Recorded here
Lines+3added
Lines−3removed
From line 1,254 where the diff opens
First seen 1 Sep 2026 this site's first read of the page
Recorded edits31to this page, all time

The whole hunk

from line 1254, old and new numbered
/
lines
from line 1254
12541254When the conversation grows too long, summarize the whole session into one user message and send only that message plus the next instruction. Nothing earlier is replayed, so there's no thinking left to fail the check, and the model reasons afresh from the summary.
12551255 
12561256<Frame>
1257 ![Simple compaction: request 4 sends the full history with thinking; request 5 sends one summary message, so nothing is checked](https://platform.claude.com/docs/images/preserved-thinking-simple-compaction.svg)
1257 ![Simple compaction: request 5 sends one user message (summary plus next instruction), so no old thinking is sent or checked](https://platform.claude.com/docs/images/preserved-thinking-simple-compaction.svg)
12581258</Frame>
12591259 
12601260```json
from line 1277
12771277The rest of this section covers a summary you write yourself.
12781278 
12791279<Frame>
1280 ![Keep-tail compaction: thinking on kept turns 3 and 4 fails; with drop\_block, the API drops it and the request succeeds](https://platform.claude.com/docs/images/preserved-thinking-keep-tail-compaction.svg)
1280 ![Keep-tail compaction: a summary of turns 1 and 2, then turns 3 to 5 verbatim. Thinking on 3 and 4 fails; drop\_block succeeds](https://platform.claude.com/docs/images/preserved-thinking-keep-tail-compaction.svg)
12811281</Frame>
12821282 
12831283Fix: keep the turns exactly as they are and send `prefix_mismatch_behavior: "drop_block"`. The API drops the stale thinking blocks, the model reads the kept turns' `text` and `tool_use` blocks, and the request succeeds.
from line 1530
15301530 </Accordion>
15311531 
15321532 <Accordion title="Does changing effort or other thinking settings between requests invalidate earlier thinking?">
1533 No. `output_config.effort`, `max_tokens`, and the `thinking` configuration aren't part of the checked prefix, which covers only `system`, `tools`, and `messages`. A top-level effort change invalidates most of the prompt cache. Where the model and platform support [per-message effort](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#effort-changes), a per-message effort change keeps the prompt cache and is used as the new effort level until changed again.
1533 No. `output_config.effort`, `max_tokens`, and the `thinking` configuration aren't part of the checked prefix, which covers only `system`, `tools`, and `messages`. A top-level effort change invalidates most of the prompt cache. Where the model and platform support per-message effort, a [per-message effort change](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#effort-changes) keeps the prompt cache and is used as the new effort level until changed again.
15341534 </Accordion>
15351535 
15361536 <Accordion title="My tool list changes mid-session. How do I avoid invalidating the conversation?">
Feedback