diff --git a/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu b/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu
index d559cda8..557efafd 100644
--- a/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu
+++ b/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu
@@ -22,6 +22,17 @@
 #include <cutlass/util/packed_stride.hpp>
 #include <cutlass/util/device_memory.h>
 
+// CUTLASS_CHECK macro (not in headers, define our own)
+#define CUTLASS_CHECK(status)                                                  \
+    do {                                                                        \
+        cutlass::Status _s = status;                                            \
+        if (_s != cutlass::Status::kSuccess) {                                  \
+            fprintf(stderr, "CUTLASS error at %s:%d: %s\n",                    \
+                    __FILE__, __LINE__, cutlassGetStatusString(_s));             \
+            return -1;                                                          \
+        }                                                                       \
+    } while (0)
+
 #if defined(CUTLASS_ARCH_MMA_SM100_SUPPORTED)
 
 using namespace cute;