diff --git a/stdlib/public/core/FloatingPointTypes.swift.gyb b/stdlib/public/core/FloatingPointTypes.swift.gyb
index 5f50ea00d5c72..98570a417fa5e 100644
--- a/stdlib/public/core/FloatingPointTypes.swift.gyb
+++ b/stdlib/public/core/FloatingPointTypes.swift.gyb
@@ -43,8 +43,11 @@ if Self == 'Float16':
     SelfDocComment = '''\
 /// A half-precision (16b), floating-point value type.
 ///
-/// `Float16` is available on Apple silicon,
-/// and unavailable on Intel when targeting macOS.'''
+/// On macOS, `Float16` is only available when targeting Apple silicon.
+/// On other supported platforms, `Float16` is available for all
+/// architectures. If the specified target supports 16-bit floating point
+/// arithmetic directly, those instructions will be used; otherwise Float16
+/// arithmetic will be emulated by the swift compiler and runtime.'''
 elif Self == 'Float':
     SelfDocComment = '''\
 /// A single-precision, floating-point value type.'''
@@ -57,9 +60,8 @@ elif Self == 'Float80':
     SelfDocComment = '''\
 /// An extended-precision, floating-point value type.
 ///
-/// `Float80` is available on Intel
-/// when the target system supports an 80-bit long double type,
-/// and unavailable on Apple silicon.'''
+/// `Float80` is available on x86 if the target system's `long double` C type
+/// is 80-bit, and unavailable otherwise.'''
 
 else:
     raise ValueError('Unhandled float type.')