diff --git a/3rdparty/amd/wheel/sglang/pyproject.toml b/3rdparty/amd/wheel/sglang/pyproject.toml
index 6f38d3952..54a04235b 100644
--- a/3rdparty/amd/wheel/sglang/pyproject.toml
+++ b/3rdparty/amd/wheel/sglang/pyproject.toml
@@ -59,7 +59,7 @@ runtime_common = [
   "timm==1.0.16",
   "torchao==0.9.0",
   "tqdm",
-  "transformers==4.57.1",
+  "transformers==5.8.1",
   "uvicorn",
   "uvloop",
   "xgrammar==0.2.0",
diff --git a/python/pyproject.toml b/python/pyproject.toml
index d5c2f3d6b..54c342596 100755
--- a/python/pyproject.toml
+++ b/python/pyproject.toml
@@ -75,7 +75,7 @@ dependencies = [
   "torchvision",
   "tqdm",
   "mistral_common>=1.11.0",
-  "transformers==5.6.0",
+  "transformers==5.8.1",
   "uvicorn",
   "uvloop",
   "watchfiles",
diff --git a/python/pyproject_cpu.toml b/python/pyproject_cpu.toml
index 963247d29..722172054 100644
--- a/python/pyproject_cpu.toml
+++ b/python/pyproject_cpu.toml
@@ -62,7 +62,7 @@ dependencies = [
   "torchvision==0.24.0",
   "tqdm",
   "mistral_common>=1.11.0",
-  "transformers==5.6.0",
+  "transformers==5.8.1",
   "triton==3.5.0",
   "uvicorn",
   "uvloop",
diff --git a/python/pyproject_npu.toml b/python/pyproject_npu.toml
index 473e9dc5c..4a0c7e728 100644
--- a/python/pyproject_npu.toml
+++ b/python/pyproject_npu.toml
@@ -59,7 +59,7 @@ dependencies = [
   "torchao==0.9.0",
   "tqdm",
   "mistral_common>=1.11.0",
-  "transformers==5.6.0",
+  "transformers==5.8.1",
   "uvicorn",
   "uvloop",
   "xgrammar==0.2.0",
diff --git a/python/pyproject_other.toml b/python/pyproject_other.toml
index c074868e4..962c10b59 100755
--- a/python/pyproject_other.toml
+++ b/python/pyproject_other.toml
@@ -59,7 +59,7 @@ runtime_common = [
   "torchao==0.9.0",
   "tqdm",
   "mistral_common>=1.11.0",
-  "transformers==5.6.0",
+  "transformers==5.8.1",
   "uvicorn",
   "uvloop",
   "xgrammar==0.2.0",
diff --git a/python/pyproject_xpu.toml b/python/pyproject_xpu.toml
index af4acb9bd..a8ee08d9e 100644
--- a/python/pyproject_xpu.toml
+++ b/python/pyproject_xpu.toml
@@ -63,7 +63,7 @@ dependencies = [
   "torchao==0.9.0+xpu",
   "tqdm",
   "mistral_common>=1.11.0",
-  "transformers==5.6.0",
+  "transformers==5.8.1",
   "uvicorn",
   "uvloop",
   # "xgrammar==0.2.0", xgrammar depends on CUDA PyTorch and Triton only