No alias env vars, no [1m], model set directly to llm-gateway/claude-sonnet-4-6.
Tests whether this avoids the clear_thinking_20251015 gateway error.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Testing whether sonnet-5 alias on the LiteLLM gateway handles
clear_thinking_20251015 without error (sonnet-4-6 does not).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…ia LiteLLM

Testing whether the codex engine + gpt-5.4-mini avoids the
clear_thinking_20251015 error that claude-sonnet-4-6/5 hit on the gateway.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…4-mini

- Bump gh-aw compiler from v0.71.5 to v0.83.1 (test branch only)
- Switch engine.model to top-level model: (deprecated in v0.83.1)
- Use api-target instead of OPENAI_BASE_URL (routes AWF proxy correctly)
- Lock file now schema v4; GH_AW_MODEL_AGENT_CODEX is properly wired

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

api-target sets the copilot proxy target, OPENAI_BASE_URL sets the
openai proxy target. Codex engine uses the openai target, so OPENAI_BASE_URL
is what redirects requests to elastic.litellm-prod.ai.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Testing haiku-4-5 while OpenAI is experiencing high demand.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>