What's new in 2.9.0 (2026-05-23)

These are the changes in inference v2.9.0.

New features

feat: add role/address to log format and structured inference logging by @m199369309 in #4898
feat: backport monitoring metrics and fix stream label / serve_count bugs by @m199369309 in #4899
feat: enhance monitoring UI with time picker, auto-refresh, and dashboard optimizations by @m199369309 in #4900
feat: add Log Center UI with Elasticsearch integration by @m199369309 in #4902
feat: add Qwen3.6-27B model support (pytorch + FP8) by @m199369309 in #4904
feat: add collapsible sidebar with persistent state by @m199369309 in #4906
feat: add page titles to Monitoring and Log Center pages by @m199369309 in #4907
feat(logs): Kibana-style enhancements for Log Center by @m199369309 in #4917
FEAT: [model] DeepSeek-V4-Flash support by @llyycchhee in #4928
feat(auth): add advanced authentication system by @m199369309 in #4931
FEAT: [model] DeepSeek-V4-Pro support by @llyycchhee in #4929
feat(i18n): update Japanese and Korean locale translations by @m199369309 in #4932
feat(web): new UI (partially implemented) by @maoyuehui in #4933
feat(vllm): support deepseek-v4 by @llyycchhee in #4938

ENH: update models JSON [llm] by @XprobeBot in #4905
ENH: update model "qwen3.6" JSON by @llyycchhee in #4923
ENH: update model "qwen3.6" JSON by @llyycchhee in #4937
ENH: update 2 models JSON ("DeepSeek-V4-Flash", "DeepSeek-V4-Pro") by @llyycchhee in #4939
BLD: Split Docker builds into separate jobs by @qinxuye in #4919

fix(ocr): replace eval() with ast.literal_eval in deepseek_ocr coordinate parsing by @Ricardo-M-L in #4872
fix: harden docs workflow against branch injection by @qinxuye in #4910
fix(ci): harden gen docs workflow by @rogercloud in #4913
fix: Fix tool_call messages handling in utils.py by @la1ty in #4915
fix(ci): update vllm to 0.21.0 for Dockerfile by @zwt-1234 in #4934

Full Changelog: v2.8.0...v2.9.0