inference/research
verifiedhigh · 0.90updated 2026-10-04

KServe v0.21.0 release notes

release-notes

Fields

id
sources/serve-kserve-release-notes
Publisher
KServe
Kind
release-notes
Published
2026-09-25
Accessed
2026-10-04
notes
Key features: direct KEDA scaling, canary traffic splitting, InferencePool for multi-model routing, NixlConnector for P/D engines, DRA (Dynamic Resource Allocation) support, OCI model delivery. Active development with 100+ PRs per release.

Notes

Key features: direct KEDA scaling, canary traffic splitting, InferencePool for multi-model routing, NixlConnector for P/D engines, DRA (Dynamic Resource Allocation) support, OCI model delivery. Active development with 100+ PRs per release.

Cross-references

This record references no other records.