fix(openclaw): switch recherche agent to deepseek-v4.1-flash and document STT quota failure

This commit is contained in:
Moritz Graf
2026-09-29 20:04:01 +02:00
parent 61ac5baa6f
commit fa6779e87d
3 changed files with 45 additions and 2 deletions
@@ -98,9 +98,11 @@ How do we architect the full-spectrum model selection across **all modalities**
2. `openrouter/qwen/qwen3.8-flash`: Provider diversity.
### 2. Autonomous Subagents (Tier 2)
* **Primary (Active Trial)**: `openrouter/z-ai/glm-5.3-flash`
* **General Subagents**: `openrouter/z-ai/glm-5.3-flash`
- **Execution**: Spawned via `sessions_spawn` with a 300-second timeout. Inherits the `tool_call` protocol guardrails from `AGENTS.md`.
* **Fallback**: `openrouter/deepseek/deepseek-v4.1-flash` for high-concurrency tool execution fallback.
* **Deep Analysis & IT Specialists (`itguy`, `recherche`)**: `openrouter/deepseek/deepseek-v4.1-flash`
- High-precision reasoning, robust nested tool loops without syntax drift, and immune to XML leakage on extensive research queries or CLI diagnostics.
* **Fallback**: `openrouter/deepseek/deepseek-v4.1-flash` / `openrouter/qwen/qwen3.8-flash`.
### 3. Background Daemon & Heartbeats (Tier 3)
* **Configuration**: `heartbeat.model: "openrouter/z-ai/glm-5.3-flash"`, `heartbeat.target: none`, `heartbeat.lightContext: true`.