Sync with Microsoft ONNX Runtime - 13062026#1135

Open

ai-fw-intg wants to merge 16 commits into

ovep-developfrom

sync_msft_13062026

Commits on Jun 11, 2026

Move OrtModelPackageApi to the experimental C API (#28990)

jambayk
and
Copilot
authored
[CUDA] Fix QMoE int4/int8 weight prepack to always use SM80 layout (#28978)

authored
webgpu: Fuse FlashAttention decode kernels and extend to any sequence length (#28389)
qjia7
authored
Fix out-of-bounds error (#28991)
apsonawane
authored
fix: Add Linux NPU discovery through sysfs accel devices (#28703)
the0cp
authored
Log EP version on inference failure and in EpDeviceUsage event, Log ORT version (#28794)

authored
Fix STFT complex input frame offsets (#28961)

GopalakrishnanN
and
Gopalakrishnan Nallasamy
authored
QMoE: fail loudly when weights_prepacked=0 but PrePack did not run (#28965)
Copilot
authored
Fix oob dereference (#29012)
apsonawane
authored

Commits on Jun 12, 2026