compaction sends mutated thinking block -> permanent 400 retry loop

Bug: compaction sends mutated `thinking` block, causes permanent retry loop (400 invalid_request_error)

opencode version: 1.16.2
Provider: anthropic, model claude-sonnet-5, extended thinking enabled (budget_tokens 8000)
Agent: compaction

What happened

During context compaction, opencode rebuilds the request history and reuses a prior assistant
turn's thinking block, but the block's thinking text field comes back empty ("thinking":"")
while the signature is preserved from the original. Anthropic's API rejects this because thinking
blocks in the last assistant message must be sent back byte-identical to how they were returned.

Because compaction retries automatically and rebuilds the same mutated payload every time, the
session gets stuck in a permanent retry loop — same 400 error, dozens of times over roughly an
hour, never recovering. Only a fresh session avoids it.

Error (from Anthropic API, relayed by opencode)

AI_APICallError
messages.N.content.1: `thinking` or `redacted_thinking` blocks in the latest assistant message
cannot be modified. These blocks must remain as they were in the original response.

N varied across retries (7, 1, 5, 3, ...) — consistent with the mutated block shifting position
as compaction repeatedly re-summarizes a growing/changing history.

Repro conditions

Long-running session that triggers auto-compaction with extended thinking enabled.
Compaction request's latest assistant message contains a thinking block whose thinking text
is "" but signature is non-empty (from the original response) — i.e. the block was stripped/
edited before resending.

Expected

Compaction should either preserve thinking blocks completely unmodified when replaying them, or
strip them entirely (not send thinking type block with empty text + stale signature) when
building the summarization request.

Additional observation

Traffic in the failing session routed through http://127.0.0.1:4097/v1/messages — a local proxy
in front of the Anthropic API — worth checking whether the proxy or opencode itself is the one
truncating the thinking block content before the request leaves the client.

This issue might be a duplicate of existing issues. Please check:

Compaction always fails on first attempt when session contains thinking blocks (Anthropic extended thinking) #31188: Same root cause — compaction with extended thinking enabled sends invalid thinking blocks, triggering the same thinking or redacted_thinking blocks in the latest assistant message cannot be modified Anthropic 400 error. That issue also includes a proposed fix (stripReasoning option in toModelMessagesEffect).

Confirming this bug affects opencode 1.18.10 with MiniMax-M3 (Anthropic-compatible endpoint), and providing empirical data.

Environment

opencode version: 1.18.10
Provider: minimax-coding-plan (uses https://api.minimax.io/anthropic/v1)
Model: MiniMax-M3 with variant: thinking
Agent: build
Session type: long-running investigation of legacy codebase

Empirical data from the affected session

Session: ses_064f372afffeAfS6ezjKK9RT5v
Title:   SIPD Convocatorias y Participaciones
Status:  STUCK in compaction loop (process still running, no progress)

Total messages:                 6,483
Total parts (DB rows):           22,837
Total parts size (serialized):   28 MB
Manual compactions performed:    6 (all "auto": false)
Auto compactions triggered:      0

Token breakdown (lifetime totals)

tokens_input:       38,391,311
tokens_output:       1,990,637
tokens_cache_read: 1,965,274,137  ← 1.96 BILLION
cost: $0.00

Behavior timeline (last 10 messages, seconds apart)

[20:47:40] assistant  in=     693  cr=    133,120  out=     115
[20:47:33] assistant  in=     431  cr=    132,864  out=      82
[20:47:24] assistant  in=     880  cr=    132,224  out=     131
[20:47:10] assistant  in=     579  cr=    131,840  out=     662
[20:46:59] assistant  in=     683  cr=    131,328  out=     386

cache_read is monotonically growing — each turn re-reads the same context.

Part distribution (proves the model is in a loop)

type=step-start     count=6058
type=step-finish    count=6045
type=tool bash      count=3438
type=text           count=3204
type=reasoning      count=1598
type=tool read      count=1126
type=tool edit      count= 922
...
type=compaction     count=   5  ← ALL manual (auto: false)

Observations consistent with #39291

auto: true in compaction config but zero auto-compactions ever fired
Last user prompt is empty — the model is in inference loop with no new user input
Each turn's cache_read keeps growing — same byte-identical context being re-sent with mutated thinking blocks
6 manual compactions previously succeeded — but the next turn after each compaction triggered the loop again
Process keeps generating messages (every ~10 seconds) but makes no progress on actual work

Compaction config (for reference)

{
  "compaction": {
    "auto": true,
    "prune": true,
    "tail_turns": 2,
    "preserve_recent_tokens": 50000,
    "reserved": 500000
  }
}

Workaround applied

kill -9 <pid> to stop the looping process
Session is unrecoverable — opening it again reproduces the same loop
New session with same task starts cleanly

Suggested mitigation for users

Until fixed, disable auto-compaction for sessions using thinking-enabled models:

{
  "compaction": {
    "auto": false   // disable auto-compaction
  }
}

Or use a non-thinking variant (variant: none or default) for long sessions.

Confirmation

This is the same bug as #39291 — auto-compaction with extended thinking enabled produces a request that the API rejects, and opencode retries it forever with the same corrupted payload. The cache_read explosion is the downstream effect of dozens of failed retries per minute over roughly an hour.

To stay organized issues are automatically closed after 60 days of no activity. If the issue is still relevant please open a new one.

Navigation Menu

Repository navigation

compaction sends mutated thinking block -> permanent 400 retry loop #39291

Bug: compaction sends mutated `thinking` block, causes permanent retry loop (400 invalid_request_error)

What happened

Error (from Anthropic API, relayed by opencode)

Repro conditions

Expected

Additional observation

Environment

Empirical data from the affected session

Token breakdown (lifetime totals)

Behavior timeline (last 10 messages, seconds apart)

Part distribution (proves the model is in a loop)

Observations consistent with #39291

Compaction config (for reference)

Workaround applied

Suggested mitigation for users

Confirmation

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

compaction sends mutated thinking block -> permanent 400 retry loop #39291

Description

Bug: compaction sends mutated thinking block, causes permanent retry loop (400 invalid_request_error)

What happened

Error (from Anthropic API, relayed by opencode)

Repro conditions

Expected

Additional observation

Activity

github-actions commented on Jul 28, 2026

iamjalberto commented on Aug 6, 2026

Environment

Empirical data from the affected session

Token breakdown (lifetime totals)

Behavior timeline (last 10 messages, seconds apart)

Part distribution (proves the model is in a loop)

Observations consistent with #39291

Compaction config (for reference)

Workaround applied

Suggested mitigation for users

Confirmation

github-actions commented on Oct 6, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Bug: compaction sends mutated `thinking` block, causes permanent retry loop (400 invalid_request_error)