diff --git a/libc/src/math/amdgpu/fmax.cpp b/libc/src/math/amdgpu/fmax.cpp
index 09624cc6f092a..09f0f942a042a 100644
--- a/libc/src/math/amdgpu/fmax.cpp
+++ b/libc/src/math/amdgpu/fmax.cpp
@@ -15,10 +15,6 @@
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(double, fmax, (double x, double y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<double>(cpp::bit_cast<uint64_t>(x) &
-                                 cpp::bit_cast<uint64_t>(y));
   return __builtin_fmax(x, y);
 }
 
diff --git a/libc/src/math/amdgpu/fmaxf.cpp b/libc/src/math/amdgpu/fmaxf.cpp
index f6ed46699a049..5913a85df6370 100644
--- a/libc/src/math/amdgpu/fmaxf.cpp
+++ b/libc/src/math/amdgpu/fmaxf.cpp
@@ -8,17 +8,11 @@
 
 #include "src/math/fmaxf.h"
 
-#include "src/__support/CPP/bit.h"
 #include "src/__support/common.h"
-#include "src/__support/macros/optimization.h"
 
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(float, fmaxf, (float x, float y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<float>(cpp::bit_cast<uint32_t>(x) &
-                                cpp::bit_cast<uint32_t>(y));
   return __builtin_fmaxf(x, y);
 }
 
diff --git a/libc/src/math/amdgpu/fmin.cpp b/libc/src/math/amdgpu/fmin.cpp
index 8977ff7a066c6..0d6f3521dcb70 100644
--- a/libc/src/math/amdgpu/fmin.cpp
+++ b/libc/src/math/amdgpu/fmin.cpp
@@ -8,17 +8,11 @@
 
 #include "src/math/fmin.h"
 
-#include "src/__support/CPP/bit.h"
 #include "src/__support/common.h"
-#include "src/__support/macros/optimization.h"
 
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(double, fmin, (double x, double y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<double>(cpp::bit_cast<uint64_t>(x) |
-                                 cpp::bit_cast<uint64_t>(y));
   return __builtin_fmin(x, y);
 }
 
diff --git a/libc/src/math/amdgpu/fminf.cpp b/libc/src/math/amdgpu/fminf.cpp
index 3be55257f6164..42744abfb3b02 100644
--- a/libc/src/math/amdgpu/fminf.cpp
+++ b/libc/src/math/amdgpu/fminf.cpp
@@ -8,17 +8,11 @@
 
 #include "src/math/fminf.h"
 
-#include "src/__support/CPP/bit.h"
 #include "src/__support/common.h"
-#include "src/__support/macros/optimization.h"
 
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(float, fminf, (float x, float y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<float>(cpp::bit_cast<uint32_t>(x) |
-                                cpp::bit_cast<uint32_t>(y));
   return __builtin_fminf(x, y);
 }
 
diff --git a/libc/src/math/nvptx/fmax.cpp b/libc/src/math/nvptx/fmax.cpp
index 09624cc6f092a..3ba65d7eccd36 100644
--- a/libc/src/math/nvptx/fmax.cpp
+++ b/libc/src/math/nvptx/fmax.cpp
@@ -8,17 +8,11 @@
 
 #include "src/math/fmax.h"
 
-#include "src/__support/CPP/bit.h"
 #include "src/__support/common.h"
-#include "src/__support/macros/optimization.h"
 
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(double, fmax, (double x, double y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<double>(cpp::bit_cast<uint64_t>(x) &
-                                 cpp::bit_cast<uint64_t>(y));
   return __builtin_fmax(x, y);
 }
 
diff --git a/libc/src/math/nvptx/fmaxf.cpp b/libc/src/math/nvptx/fmaxf.cpp
index f6ed46699a049..e977082b39f40 100644
--- a/libc/src/math/nvptx/fmaxf.cpp
+++ b/libc/src/math/nvptx/fmaxf.cpp
@@ -15,10 +15,6 @@
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(float, fmaxf, (float x, float y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<float>(cpp::bit_cast<uint32_t>(x) &
-                                cpp::bit_cast<uint32_t>(y));
   return __builtin_fmaxf(x, y);
 }
 
diff --git a/libc/src/math/nvptx/fmin.cpp b/libc/src/math/nvptx/fmin.cpp
index 8977ff7a066c6..0d6f3521dcb70 100644
--- a/libc/src/math/nvptx/fmin.cpp
+++ b/libc/src/math/nvptx/fmin.cpp
@@ -8,17 +8,11 @@
 
 #include "src/math/fmin.h"
 
-#include "src/__support/CPP/bit.h"
 #include "src/__support/common.h"
-#include "src/__support/macros/optimization.h"
 
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(double, fmin, (double x, double y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<double>(cpp::bit_cast<uint64_t>(x) |
-                                 cpp::bit_cast<uint64_t>(y));
   return __builtin_fmin(x, y);
 }
 
diff --git a/libc/src/math/nvptx/fminf.cpp b/libc/src/math/nvptx/fminf.cpp
index 3be55257f6164..42744abfb3b02 100644
--- a/libc/src/math/nvptx/fminf.cpp
+++ b/libc/src/math/nvptx/fminf.cpp
@@ -8,17 +8,11 @@
 
 #include "src/math/fminf.h"
 
-#include "src/__support/CPP/bit.h"
 #include "src/__support/common.h"
-#include "src/__support/macros/optimization.h"
 
 namespace LIBC_NAMESPACE {
 
 LLVM_LIBC_FUNCTION(float, fminf, (float x, float y)) {
-  // FIXME: The builtin function does not correctly handle the +/-0.0 case.
-  if (LIBC_UNLIKELY(x == y))
-    return cpp::bit_cast<float>(cpp::bit_cast<uint32_t>(x) |
-                                cpp::bit_cast<uint32_t>(y));
   return __builtin_fminf(x, y);
 }