viktor/infra
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

openclaw: revert model swap + document codex re-auth path

Browse source 
The previous commit promoted modelrelay/auto-fastest to primary as a
workaround for the expired openai-codex OAuth token. But modelrelay
routes to small tool-call-shy models (nvidia/stepfun-ai/step-3.5-flash)
that hallucinate answers instead of using ssh / curl / etc. — exactly
what the v4 learning loop is supposed to leverage.

Revert primary back to openai-codex/gpt-5.4-mini (gpt-5.4-mini is the
only mini variant the Codex backend accepts for ChatGPT Plus tier),
and inline the re-auth command in the model-block comment so future
sessions know exactly what to do when the OAuth token expires:

  kubectl -n openclaw exec -it $(kubectl -n openclaw get pods \
    -l app=openclaw -o jsonpath='{.items[0].metadata.name}') \
    -c openclaw -- node /app/openclaw.mjs models auth login \
    --provider openai-codex

modelrelay/auto-fastest stays in the fallback chain so the agent
remains partially usable while the token is expired.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
Viktor Barzin 2026-05-22 14:12:30 +00:00
parent 6457aa6d8f
commit d4c76a07a2
1 changed files with 17 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

22
stacks/openclaw/main.tf
View file

@ -132,12 +132,24 @@ resource "kubernetes_config_map" "openclaw_config" {
mode = "off"
}
model = {
# ChatGPT Plus OAuth via openai-codex plugin (account: ancaelena98@gmail.com).
# gpt-5.4-mini is the only mini variant the Codex backend accepts for Plus tier;
# gpt-5-mini / gpt-5.1-codex-mini return model_not_found / "not supported with
# ChatGPT account". Plus rate-card: 1,200–7,000 local msgs / 5h on gpt-5.4-mini.
# ChatGPT Plus OAuth via openai-codex plugin (account:
# ancaelena98@gmail.com). gpt-5.4-mini is the only mini
# variant the Codex backend accepts for Plus tier;
# gpt-5-mini / gpt-5.1-codex-mini return model_not_found
# / "not supported with ChatGPT account". Plus rate-card:
# 1,200–7,000 local msgs / 5h on gpt-5.4-mini.
#
# If you see "No API key found for provider openai-codex"
# / "OAuth refresh failed" in logs, the OAuth token has
# expired. Re-auth:
# kubectl -n openclaw exec -it $(kubectl -n openclaw \
# get pods -l app=openclaw -o jsonpath='{.items[0].metadata.name}') \
# -c openclaw -- node /app/openclaw.mjs models auth login \
# --provider openai-codex
# Follow the OAuth URL+code prompt. Tokens persist on the
# openclaw-home PVC so it sticks across pod restarts.
primary = "openai-codex/gpt-5.4-mini"
fallbacks = ["openai-codex/gpt-5.5", "nim/qwen/qwen3-coder-480b-a35b-instruct", "modelrelay/auto-fastest"]
fallbacks = ["openai-codex/gpt-5.5", "modelrelay/auto-fastest", "nim/qwen/qwen3-coder-480b-a35b-instruct"]
}
models = {
"modelrelay/auto-fastest" = {}