diff --git a/gen/gen_code.cpp b/gen/gen_code.cpp
index 8a6f7ac..251351c 100644
--- a/gen/gen_code.cpp
+++ b/gen/gen_code.cpp
@@ -1733,7 +1733,7 @@ void put()
 			uint8_t code;
 		} tbl[] = {
 			{ "andn", T_0F38 | T_VEX | T_NF, 0xF2 },
-			{ "mulx", T_F2 | T_0F38, 0xF6 },
+			{ "mulx", T_F2 | T_0F38 | T_VEX , 0xF6 },
 			{ "pdep", T_F2 | T_0F38, 0xF5 },
 			{ "pext", T_F3 | T_0F38, 0xF5 },
 		};
diff --git a/test/apx.cpp b/test/apx.cpp
index d07c6cc..633b4bc 100644
--- a/test/apx.cpp
+++ b/test/apx.cpp
@@ -457,7 +457,7 @@ CYBOZU_TEST_AUTO(NF)
 	CYBOZU_TEST_EQUAL_ARRAY(c.getCode(), tbl, n);
 }
 
-CYBOZU_TEST_AUTO(andn)
+CYBOZU_TEST_AUTO(andn_etc)
 {
 	struct Code : Xbyak::CodeGenerator {
 		Code()
@@ -467,11 +467,21 @@ CYBOZU_TEST_AUTO(andn)
 			andn(eax, ecx, r8d); // 32bit
 			andn(r29, r30, r31); // 64bit
 			andn(r29, r30, ptr [r31+r20*4]);
+
+			mulx(rax, rcx, r8);
+			mulx(eax, ecx, r8d);
+			mulx(r29, r30, r31);
+			mulx(r29, r30, ptr [r31+r20*4]);
 		}
 	} c;
 	const uint8_t tbl[] = {
+		// andn
 		0xc4, 0xc2, 0xf0, 0xf2, 0xc0, 0x62, 0xd2, 0xf4, 0x08, 0xf2, 0xc0, 0xc4, 0xc2, 0x70, 0xf2, 0xc0,
 		0x62, 0x4a, 0x8c, 0x00, 0xf2, 0xef, 0x62, 0x4a, 0x88, 0x00, 0xf2, 0x2c, 0xa7,
+
+		// mulx
+		0xc4, 0xc2, 0xf3, 0xf6, 0xc0, 0xc4, 0xc2, 0x73, 0xf6, 0xc0, 0x62, 0x4a, 0x8f, 0x00, 0xf6, 0xef,
+0x62, 0x4a, 0x8b, 0x00, 0xf6, 0x2c, 0xa7,
 	};
 	const size_t n = sizeof(tbl);
 	CYBOZU_TEST_EQUAL(c.getSize(), n);
diff --git a/xbyak/xbyak_mnemonic.h b/xbyak/xbyak_mnemonic.h
index 5d31a63..4f769f0 100644
--- a/xbyak/xbyak_mnemonic.h
+++ b/xbyak/xbyak_mnemonic.h
@@ -560,7 +560,7 @@ void mulpd(const Xmm& xmm, const Operand& op) { opSSE(xmm, op, T_0F | T_66, 0x59
 void mulps(const Xmm& xmm, const Operand& op) { opSSE(xmm, op, T_0F, 0x59, isXMM_XMMorMEM); }
 void mulsd(const Xmm& xmm, const Operand& op) { opSSE(xmm, op, T_0F | T_F2, 0x59, isXMM_XMMorMEM); }
 void mulss(const Xmm& xmm, const Operand& op) { opSSE(xmm, op, T_0F | T_F3, 0x59, isXMM_XMMorMEM); }
-void mulx(const Reg& r1, const Reg32e& r2, const Operand& op) { opRRO(r1, r2, op, T_F2|T_0F38, 0xf6); }
+void mulx(const Reg& r1, const Reg32e& r2, const Operand& op) { opRRO(r1, r2, op, T_VEX|T_F2|T_0F38, 0xf6); }
 void mwait() { db(0x0F); db(0x01); db(0xC9); }
 void mwaitx() { db(0x0F); db(0x01); db(0xFB); }
 void neg(const Operand& op) { opRext(op, 0, 3, 0, 0xF6); }