Skip to content
CC Switch WikiUnofficial Community Wiki

Reasoning Level Not Adapted - CC Switch

Quick fix

Re-add the OpenCode Go preset after upgrading to get the correct reasoning_effort config.

Symptom

Section titled “Symptom”
Error output
codex中模型选择正常，但思考等级未适配
English translation

Model selection in Codex works normally, but the reasoning level is not adapted

Cause

Section titled “Cause”

The OpenCode Go preset routes through the Zen gateway of opencode.ai. Older versions of cc-switch incorrectly used the ZhiPu official interface's thinking:{type} parameter format, whereas the Zen gateway only accepts the top-level reasoning_effort field (with values low/medium/high/max). This caused the model to call successfully but failed to pass the reasoning level correctly.

Additionally, although the platform-level parameter sending logic was fixed and applied automatically, existing supplier configuration files lacked a per-model valid gear table (modelCatalog[].reasoningLevels). Therefore, simply upgrading the software does not automatically fix the reasoning level adaptation for existing suppliers; you must re-add this preset supplier to load the complete gear mapping configuration.

Fix

Section titled “Fix”

  1. Upgrade cc-switch to the latest version containing PR #6123 fix.

  2. In the supplier management interface, delete the existing OpenCode Go preset supplier, then re-add it once. This ensures that settings_config contains the correct per-model gear table.

    # 操作位置：cc-switch 图形界面的供应商编辑/添加页面

Affected Versions

Section titled “Affected Versions”
ToolCodex
Version3.19.1
PlatformsWindows

Source Issues

Section titled “Source Issues”

This page is distilled from 2 real issues

FAQ

Section titled “FAQ”
Why is my OpenCode Go supplier's reasoning level still unresponsive after upgrading?
Because the upgrade only fixes the underlying parameter sending logic but does not automatically backfill the per-model gear table for existing suppliers. You need to delete and re-add the OpenCode Go preset supplier to obtain the complete configuration.
What reasoning levels does the Zen gateway support?
The Zen gateway supports four levels: low, medium, high, max. cc-switch maps Codex levels such as minimal/xhigh to these values. Specifically which levels are effective for a certain model (e.g., glm-5.2 may only have high/max) is validated by the gateway side.
Section titled “Related problems”
The model cannot discover any tools - CodexUpgrade cc-switch past 3.16.1 and edit cc-switch-model-catalog.json to set supports_search_tool to false for the DeepSeek models.2 issues mergedUnknown name "prompt_cache_key" - cc-switchUpgrade cc-switch to v3.16.2+ to fix prompt_cache_key injection, or use an OpenRouter endpoint that natively speaks the Claude Code format.6 issues merged401 Unauthorized: Incorrect API key providedAfter switching to OpenAI Official or turning the proxy off, clear the leftover third-party OPENAI_API_KEY from ~/.codex/auth.json and restart Codex.2 issues merged

This is an unofficial community wiki with no affiliation to the cc-switch authors or the project itself. Its content is compiled from the project's public GitHub issues. This site distributes no software.