diff --git a/include/cglm/simd/avx/affine.h b/include/cglm/simd/avx/affine.h
index 369bcf7..b02ff0c 100644
--- a/include/cglm/simd/avx/affine.h
+++ b/include/cglm/simd/avx/affine.h
@@ -29,20 +29,22 @@ glm_mul_avx(mat4 m1, mat4 m2, mat4 dest) {
 
   /* 0x03: 0b00000011 */
   y4 = _mm256_permute2f128_ps(y2, y2, 0x03); /* d c b a h g f e */
-  y5 = _mm256_permute2f128_ps(y3, y3, 0x03); /* l k j i l k j i */
+  y5 = _mm256_permute2f128_ps(y3, y3, 0x03); /* l k j i p o n m */
 
   /* f f f f a a a a */
-  /* g g g g c c c c */
+  /* h h h h c c c c */
   /* e e e e b b b b */
-  y7 = _mm256_permute_ps(y0, 0xAA); /* 0xAA: 0b10101010 */
+  /* g g g g d d d d */
   y6 = _mm256_permutevar_ps(y0, _mm256_set_epi32(1, 1, 1, 1, 0, 0, 0, 0));
+  y7 = _mm256_permutevar_ps(y0, _mm256_set_epi32(3, 3, 3, 3, 2, 2, 2, 2));
   y8 = _mm256_permutevar_ps(y0, _mm256_set_epi32(0, 0, 0, 0, 1, 1, 1, 1));
+  y9 = _mm256_permutevar_ps(y0, _mm256_set_epi32(2, 2, 2, 2, 3, 3, 3, 3));
 
   glmm_store256(dest[0],
                 _mm256_add_ps(_mm256_add_ps(_mm256_mul_ps(y2, y6),
-                                            _mm256_mul_ps(y4, y8)),
-                              _mm256_mul_ps(y5, y7)));
-
+                                            _mm256_mul_ps(y3, y7)),
+                              _mm256_add_ps(_mm256_mul_ps(y4, y8),
+                                            _mm256_mul_ps(y5, y9))));
 
   /* n n n n i i i i */
   /* p p p p k k k k */