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

openclaw: switch primary model to qwen3-coder-480b (qwen3.5-397b dead on NIM)

Browse source 
NVIDIA retired nim/qwen/qwen3.5-397b-a17b — modelrelay shows consistent
TIMEOUTs over 24h+ of pings, and nim/nvidia/llama-3.1-nemotron-ultra-253b-v1
returns 404. With both gone the openclaw failover never reached
mistral-large-3 in time, so every message hung until the 120s embedded-run
timeout. Promote qwen3-coder-480b-a35b-instruct (already in models list, UP
~1-2s, 256k ctx) to primary; drop the dead nemotron-ultra fallback.
This commit is contained in:
Viktor Barzin 2026-05-06 20:35:38 +00:00
parent f006b48566
commit 8b180f7662
1 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
stacks/openclaw/main.tf
View file

@ -131,8 +131,8 @@ resource "kubernetes_config_map" "openclaw_config" {
mode = "off" mode = "off"
} }
model = { model = {
primary = "nim/qwen/qwen3.5-397b-a17b" primary = "nim/qwen/qwen3-coder-480b-a35b-instruct"
fallbacks = ["nim/mistralai/mistral-large-3-675b-instruct-2512", "nim/nvidia/llama-3.1-nemotron-ultra-253b-v1", "modelrelay/auto-fastest"] fallbacks = ["nim/mistralai/mistral-large-3-675b-instruct-2512", "modelrelay/auto-fastest"]
} }
models = { models = {
"modelrelay/auto-fastest" = {} "modelrelay/auto-fastest" = {}