feat(channel): add Baichuan type + fix max_tokens & Azure AI Foundry /openai/v1 - #6597

Open

aiastia wants to merge 9 commits into

QuantumNous:mainfrom

aiastia:feat/baichuan-channel

Commits on Aug 2, 2026

feat(channel): add dedicated Baichuan channel type
aiastia
committed
c386058
fix(channel-test): use max_completion_tokens for models that reject max_tokens
aiastia
committed
a4c4068
fix(azure): support AI Foundry /openai/v1 endpoint in Azure channel type
aiastia
committed
30c696c
ci: add GHCR build workflow for fork self-hosting
aiastia
committed
3f69290
ci: fix GHCR workflow — base on official docker-image-branch.yml
aiastia
committed
2921aa9
fix(azure): avoid /openai prefix duplication with /openai/v1 endpoint
aiastia
committed
eacc213
fix(azure): detect AI Foundry by host, not by request path
aiastia
committed
c2f5968
fix(azure): detect AI Foundry by /openai suffix in base URL
aiastia
committed
1f0c83e

Commits on Aug 3, 2026

fix(cohere): parse {models:[{name}]} when fetching upstream model list
aiastia
committed
a777034