diff --git a/libc/test/src/stdlib/StrtolTest.h b/libc/test/src/stdlib/StrtolTest.h
--- a/libc/test/src/stdlib/StrtolTest.h
+++ b/libc/test/src/stdlib/StrtolTest.h
@@ -8,6 +8,7 @@
 
 #include "src/__support/CPP/limits.h"
 #include "src/__support/CPP/type_traits.h"
+#include "src/__support/macros/properties/architectures.h"
 #include "src/errno/libc_errno.h"
 #include "test/UnitTest/Test.h"
 
@@ -198,6 +199,12 @@
   }
 
   void DecodeInOtherBases(FunctionT func) {
+#if defined(LIBC_TARGET_ARCH_IS_GPU)
+    // This test is excessively slow on the GPU so we limit the nested depth.
+    constexpr int limit = 1;
+#else
+    constexpr int limit = 36;
+#endif
     char small_string[4] = {'\0', '\0', '\0', '\0'};
     for (int base = 2; base <= 36; ++base) {
       for (int first_digit = 0; first_digit <= 36; ++first_digit) {
@@ -218,7 +225,7 @@
     for (int base = 2; base <= 36; ++base) {
       for (int first_digit = 0; first_digit <= 36; ++first_digit) {
         small_string[0] = int_to_b36_char(first_digit);
-        for (int second_digit = 0; second_digit <= 36; ++second_digit) {
+        for (int second_digit = 0; second_digit <= limit; ++second_digit) {
           small_string[1] = int_to_b36_char(second_digit);
           if (first_digit < base && second_digit < base) {
             libc_errno = 0;
@@ -243,9 +250,9 @@
     for (int base = 2; base <= 36; ++base) {
       for (int first_digit = 0; first_digit <= 36; ++first_digit) {
         small_string[0] = int_to_b36_char(first_digit);
-        for (int second_digit = 0; second_digit <= 36; ++second_digit) {
+        for (int second_digit = 0; second_digit <= limit; ++second_digit) {
           small_string[1] = int_to_b36_char(second_digit);
-          for (int third_digit = 0; third_digit <= 36; ++third_digit) {
+          for (int third_digit = 0; third_digit <= limit; ++third_digit) {
             small_string[2] = int_to_b36_char(third_digit);
 
             if (first_digit < base && second_digit < base &&