docs(models): clarify moonshot thinking and failover stop-reason errors

This commit is contained in:
Peter Steinberger
2026-03-03 01:11:18 +00:00
parent 17b40c4a59
commit 42626648d7
2 changed files with 4 additions and 0 deletions

View File

@@ -83,6 +83,9 @@ When a profile fails due to auth/ratelimit errors (or a timeout that looks
like rate limiting), OpenClaw marks it in cooldown and moves to the next profile. like rate limiting), OpenClaw marks it in cooldown and moves to the next profile.
Format/invalidrequest errors (for example Cloud Code Assist tool call ID Format/invalidrequest errors (for example Cloud Code Assist tool call ID
validation failures) are treated as failoverworthy and use the same cooldowns. validation failures) are treated as failoverworthy and use the same cooldowns.
OpenAI-compatible stop-reason errors such as `Unhandled stop reason: error`,
`stop reason: error`, and `reason: error` are classified as timeout/failover
signals.
Cooldowns use exponential backoff: Cooldowns use exponential backoff:

View File

@@ -22,6 +22,7 @@ title: "Thinking Levels"
- Provider notes: - Provider notes:
- Anthropic Claude 4.6 models default to `adaptive` when no explicit thinking level is set. - Anthropic Claude 4.6 models default to `adaptive` when no explicit thinking level is set.
- Z.AI (`zai/*`) only supports binary thinking (`on`/`off`). Any non-`off` level is treated as `on` (mapped to `low`). - Z.AI (`zai/*`) only supports binary thinking (`on`/`off`). Any non-`off` level is treated as `on` (mapped to `low`).
- Moonshot (`moonshot/*`) maps `/think off` to `thinking: { type: "disabled" }` and any non-`off` level to `thinking: { type: "enabled" }`. When thinking is enabled, Moonshot only accepts `tool_choice` `auto|none`; OpenClaw normalizes incompatible values to `auto`.
## Resolution order ## Resolution order