diff --git a/src/aro/Attribute.zig b/src/aro/Attribute.zig
index a09bf3b0..6f46b72b 100644
--- a/src/aro/Attribute.zig
+++ b/src/aro/Attribute.zig
@@ -904,7 +904,7 @@ pub fn applyFunctionAttributes(p: *Parser, ty: Type, attr_buf_start: usize) !Typ
                 else => try p.errStr(.callconv_not_supported, tok, p.tok_ids[tok].lexeme().?),
             },
             .vectorcall => switch (p.comp.target.cpu.arch) {
-                .x86, .aarch64, .aarch64_be, .aarch64_32 => try p.attr_application_buf.append(p.gpa, attr),
+                .x86, .aarch64, .aarch64_be => try p.attr_application_buf.append(p.gpa, attr),
                 else => try p.errStr(.callconv_not_supported, tok, p.tok_ids[tok].lexeme().?),
             },
         },
diff --git a/src/aro/Compilation.zig b/src/aro/Compilation.zig
index 12ccc855..fc8e1621 100644
--- a/src/aro/Compilation.zig
+++ b/src/aro/Compilation.zig
@@ -666,7 +666,7 @@ fn generateBuiltinTypes(comp: *Compilation) !void {
         .arm, .armeb, .thumb, .thumbeb => .{
             .specifier = if (os != .windows and os != .netbsd and os != .openbsd) .uint else .int,
         },
-        .aarch64, .aarch64_be, .aarch64_32 => .{
+        .aarch64, .aarch64_be => .{
             .specifier = if (!os.isDarwin() and os != .netbsd) .uint else .int,
         },
         .x86_64, .x86 => .{ .specifier = if (os == .windows) .ushort else .int },
diff --git a/src/aro/target.zig b/src/aro/target.zig
index 81ba4881..ddb9eb42 100644
--- a/src/aro/target.zig
+++ b/src/aro/target.zig
@@ -137,7 +137,7 @@ pub fn float80Type(target: std.Target) ?Type {
 pub fn defaultFunctionAlignment(target: std.Target) u8 {
     return switch (target.cpu.arch) {
         .arm, .armeb => 4,
-        .aarch64, .aarch64_32, .aarch64_be => 4,
+        .aarch64, .aarch64_be => 4,
         .sparc, .sparcel, .sparc64 => 4,
         .riscv64 => 2,
         else => 1,
@@ -327,7 +327,6 @@ pub const FPSemantics = enum {
     pub fn halfPrecisionType(target: std.Target) ?FPSemantics {
         switch (target.cpu.arch) {
             .aarch64,
-            .aarch64_32,
             .aarch64_be,
             .arm,
             .armeb,
@@ -483,7 +482,6 @@ pub fn get32BitArchVariant(target: std.Target) ?std.Target {
         .kalimba,
         .lanai,
         .wasm32,
-        .aarch64_32,
         .spirv,
         .spirv32,
         .loongarch32,
@@ -548,7 +546,6 @@ pub fn get64BitArchVariant(target: std.Target) ?std.Target {
         .x86_64,
         => {}, // Already 64 bit
 
-        .aarch64_32 => copy.cpu.arch = .aarch64,
         .arm => copy.cpu.arch = .aarch64,
         .armeb => copy.cpu.arch = .aarch64_be,
         .loongarch32 => copy.cpu.arch = .loongarch64,
@@ -583,7 +580,6 @@ pub fn toLLVMTriple(target: std.Target, buf: []u8) []const u8 {
         .armeb => "armeb",
         .aarch64 => "aarch64",
         .aarch64_be => "aarch64_be",
-        .aarch64_32 => "aarch64_32",
         .arc => "arc",
         .avr => "avr",
         .bpfel => "bpfel",
@@ -696,7 +692,7 @@ pub fn toLLVMTriple(target: std.Target, buf: []u8) []const u8 {
     writer.writeByte('-') catch unreachable;
 
     const llvm_abi = switch (target.abi) {
-        .none => "unknown",
+        .none, .ilp32 => "unknown",
         .gnu => "gnu",
         .gnuabin32 => "gnuabin32",
         .gnuabi64 => "gnuabi64",