From 678baf53724e06cef7f07628ae8fd4cc6c96f11a Mon Sep 17 00:00:00 2001
From: Wei Zhao <51183510+wzhao18@users.noreply.github.com>
Date: Tue, 29 Sep 2026 22:34:36 -0400
Subject: [PATCH] [Dependency] Upgrade FlashInfer to 0.7.0.post1 (#59323)

Signed-off-by: wzhao18 <wzhao18.sz@gmail.com>
---
 docker/Dockerfile                 | 2 +-
 docker/versions.json              | 2 +-
 requirements/cuda.txt             | 4 ++--
 requirements/rubin-prerelease.txt | 6 +++---
 4 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/docker/Dockerfile b/docker/Dockerfile
index 5f4c4ab35f..ba046adae1 100644
--- a/docker/Dockerfile
+++ b/docker/Dockerfile
@@ -1004,7 +1004,7 @@ RUN --mount=type=cache,target=/opt/uv/cache \
 # Install FlashInfer JIT cache (requires CUDA-version-specific index URL)
 # https://docs.flashinfer.ai/installation.html
 # From versions.json: .flashinfer.version
-ARG FLASHINFER_VERSION=0.7.0
+ARG FLASHINFER_VERSION=0.7.0.post1
 RUN --mount=type=cache,target=/opt/uv/cache \
     if [ "${INSTALL_RUBIN_PRERELEASE}" != "true" ]; then \
         uv pip install --system flashinfer-jit-cache==${FLASHINFER_VERSION} \
diff --git a/docker/versions.json b/docker/versions.json
index ec23b95186..bcec27fd55 100644
--- a/docker/versions.json
+++ b/docker/versions.json
@@ -71,7 +71,7 @@
       "default": "true"
     },
     "FLASHINFER_VERSION": {
-      "default": "0.7.0"
+      "default": "0.7.0.post1"
     },
     "GDRCOPY_CUDA_VERSION": {
       "default": "12.8"
diff --git a/requirements/cuda.txt b/requirements/cuda.txt
index 2bbf1e244d..5cd1e717be 100644
--- a/requirements/cuda.txt
+++ b/requirements/cuda.txt
@@ -14,8 +14,8 @@ PyNvVideoCodec==2.0.4
 # flashinfer-cubin is not on PyPI since 0.6.14; setup.py excludes it from
 # install_requires so the published wheel does not carry an unresolvable pin
 --extra-index-url https://flashinfer.ai/whl/
-flashinfer-python==0.7.0
-flashinfer-cubin==0.7.0
+flashinfer-python==0.7.0.post1
+flashinfer-cubin==0.7.0.post1
 apache-tvm-ffi==0.1.11
 tilelang==0.1.12
 nvidia-cudnn-frontend>=1.19.1
diff --git a/requirements/rubin-prerelease.txt b/requirements/rubin-prerelease.txt
index d02fb38884..e5915b6d13 100644
--- a/requirements/rubin-prerelease.txt
+++ b/requirements/rubin-prerelease.txt
@@ -16,11 +16,11 @@ torchaudio==2.11.0.dev20260911+cu134
 # INSTALL_RUBIN_PRERELEASE=true, so they must be pinned here instead to
 # still get installed.
 nvidia-cutlass-dsl[cu13]==4.8.0.dev0
-flashinfer-python==0.7.0
+flashinfer-python==0.7.0.post1
 quack-kernels==0.6.5
 
 cuda-python==13.4.1
 cuda-bindings==13.4.1
-flashinfer-jit-cache==0.7.0+cu134
-flashinfer-cubin==0.7.0
+flashinfer-jit-cache==0.7.0.post1+cu134
+flashinfer-cubin==0.7.0.post1
 nixl==1.4.1