diff --git a/src/resources/cache/hashes.json b/src/resources/cache/hashes.json
index ea368705c448..e0459422560f 100644
--- a/src/resources/cache/hashes.json
+++ b/src/resources/cache/hashes.json
@@ -97,7 +97,7 @@
   "webgpu/shader/execution/unary/f16_arithmetic.bin": "4a20db6d",
   "webgpu/shader/execution/unary/f16_conversion.bin": "31f72f5a",
   "webgpu/shader/execution/unary/f32_arithmetic.bin": "f1c311cb",
-  "webgpu/shader/execution/unary/f32_conversion.bin": "57dc324c",
+  "webgpu/shader/execution/unary/f32_conversion.bin": "7539cdb3",
   "webgpu/shader/execution/unary/i32_arithmetic.bin": "de945eec",
   "webgpu/shader/execution/unary/i32_conversion.bin": "1728a03e",
   "webgpu/shader/execution/unary/u32_conversion.bin": "9e6ca0ce",
diff --git a/src/resources/cache/webgpu/shader/execution/unary/f32_conversion.bin b/src/resources/cache/webgpu/shader/execution/unary/f32_conversion.bin
index bdcc0c72988e..66b2bc73f889 100644
Binary files a/src/resources/cache/webgpu/shader/execution/unary/f32_conversion.bin and b/src/resources/cache/webgpu/shader/execution/unary/f32_conversion.bin differ
diff --git a/src/webgpu/listing_meta.json b/src/webgpu/listing_meta.json
index 878a7155697e..2360a199e3e4 100644
--- a/src/webgpu/listing_meta.json
+++ b/src/webgpu/listing_meta.json
@@ -1681,6 +1681,9 @@
   "webgpu:shader,execution,expression,unary,f16_conversion:i32:*": { "subcaseMS": 24.557 },
   "webgpu:shader,execution,expression,unary,f16_conversion:u32:*": { "subcaseMS": 84.500 },
   "webgpu:shader,execution,expression,unary,f32_arithmetic:negation:*": { "subcaseMS": 16.400 },
+  "webgpu:shader,execution,expression,unary,f32_conversion:abstract_float:*": { "subcaseMS": 688.718 },
+  "webgpu:shader,execution,expression,unary,f32_conversion:abstract_float_mat:*": { "subcaseMS": 1409.951 },
+  "webgpu:shader,execution,expression,unary,f32_conversion:abstract_int:*": { "subcaseMS": 506.131 },
   "webgpu:shader,execution,expression,unary,f32_conversion:bool:*": { "subcaseMS": 7.182 },
   "webgpu:shader,execution,expression,unary,f32_conversion:f16:*": { "subcaseMS": 107.463 },
   "webgpu:shader,execution,expression,unary,f32_conversion:f16_mat:*": { "subcaseMS": 60.170 },
diff --git a/src/webgpu/shader/execution/expression/unary/f32_conversion.cache.ts b/src/webgpu/shader/execution/expression/unary/f32_conversion.cache.ts
index f61435f07ce4..8ce39eccfd1b 100644
--- a/src/webgpu/shader/execution/expression/unary/f32_conversion.cache.ts
+++ b/src/webgpu/shader/execution/expression/unary/f32_conversion.cache.ts
@@ -1,7 +1,8 @@
-import { bool, f16, f32, i32, u32 } from '../../../../util/conversion.js';
-import { FP } from '../../../../util/floating_point.js';
+import { abstractInt, bool, f16, f32, i32, u32 } from '../../../../util/conversion.js';
+import { FP, FPInterval } from '../../../../util/floating_point.js';
 import {
   fullI32Range,
+  fullI64Range,
   fullU32Range,
   scalarF16Range,
   scalarF32Range,
@@ -10,6 +11,12 @@ import {
 } from '../../../../util/math.js';
 import { makeCaseCache } from '../case_cache.js';
 
+const f32FiniteRangeInterval = new FPInterval(
+  'f32',
+  FP.f32.constants().negative.min,
+  FP.f32.constants().positive.max
+);
+
 // Cases: f32_matCxR_[non_]const
 const f32_mat_cases = ([2, 3, 4] as const)
   .flatMap(cols =>
@@ -46,6 +53,23 @@ const f16_mat_cases = ([2, 3, 4] as const)
   )
   .reduce((a, b) => ({ ...a, ...b }), {});
 
+// Cases: abstract_float_matCxR
+// Note that abstract float values may be not exactly representable in f32
+// and/or out of range.
+const abstract_float_mat_cases = ([2, 3, 4] as const)
+  .flatMap(cols =>
+    ([2, 3, 4] as const).map(rows => ({
+      [`abstract_float_mat${cols}x${rows}`]: () => {
+        return FP.abstract.generateMatrixToMatrixCases(
+          FP.abstract.sparseMatrixRange(cols, rows),
+          'finite',
+          FP.f32.correctlyRoundedMatrix
+        );
+      },
+    }))
+  )
+  .reduce((a, b) => ({ ...a, ...b }), {});
+
 export const d = makeCaseCache('unary/f32_conversion', {
   bool: () => {
     return [
@@ -63,6 +87,13 @@ export const d = makeCaseCache('unary/f32_conversion', {
       return { input: i32(i), expected: FP.f32.correctlyRoundedInterval(i) };
     });
   },
+  abstract_int: () => {
+    return [...fullI64Range()]
+      .filter(v => f32FiniteRangeInterval.contains(Number(v)))
+      .map(i => {
+        return { input: abstractInt(i), expected: FP.f32.correctlyRoundedInterval(Number(i)) };
+      });
+  },
   f32: () => {
     return scalarF32Range().map(f => {
       return { input: f32(f), expected: FP.f32.correctlyRoundedInterval(f) };
@@ -74,6 +105,15 @@ export const d = makeCaseCache('unary/f32_conversion', {
       return { input: f16(f), expected: FP.f32.correctlyRoundedInterval(f) };
     });
   },
+  // Note that abstract float values may be not exactly representable in f32.
+  abstract_float: () => {
+    return FP.abstract.generateScalarToIntervalCases(
+      [...FP.abstract.scalarRange()],
+      'finite',
+      FP.f32.correctlyRoundedInterval
+    );
+  },
   ...f32_mat_cases,
   ...f16_mat_cases,
+  ...abstract_float_mat_cases,
 });
diff --git a/src/webgpu/shader/execution/expression/unary/f32_conversion.spec.ts b/src/webgpu/shader/execution/expression/unary/f32_conversion.spec.ts
index 464fdee44e79..7a42334cb860 100644
--- a/src/webgpu/shader/execution/expression/unary/f32_conversion.spec.ts
+++ b/src/webgpu/shader/execution/expression/unary/f32_conversion.spec.ts
@@ -5,7 +5,7 @@ Execution Tests for the f32 conversion operations
 import { makeTestGroup } from '../../../../../common/framework/test_group.js';
 import { GPUTest } from '../../../../gpu_test.js';
 import { Type } from '../../../../util/conversion.js';
-import { ShaderBuilder, allInputSources, run } from '../expression.js';
+import { ShaderBuilder, allInputSources, run, onlyConstInputSource } from '../expression.js';
 
 import { d } from './f32_conversion.cache.js';
 import { unary } from './unary.js';
@@ -73,6 +73,32 @@ Converted to f32
     await run(t, vectorizeToExpression(t.params.vectorize), [Type.i32], Type.f32, t.params, cases);
   });
 
+g.test('abstract_int')
+  .specURL('https://www.w3.org/TR/WGSL/#value-constructor-builtin-function')
+  .desc(
+    `
+f32(e), where e is an AbstractInt
+
+Converted to f32, +/-Inf if out of range
+`
+  )
+  .params(u =>
+    u
+      .combine('inputSource', onlyConstInputSource)
+      .combine('vectorize', [undefined, 2, 3, 4] as const)
+  )
+  .fn(async t => {
+    const cases = await d.get('abstract_int');
+    await run(
+      t,
+      vectorizeToExpression(t.params.vectorize),
+      [Type.abstractInt],
+      Type.f32,
+      t.params,
+      cases
+    );
+  });
+
 g.test('f32')
   .specURL('https://www.w3.org/TR/WGSL/#value-constructor-builtin-function')
   .desc(
@@ -166,3 +192,52 @@ g.test('f16_mat')
       cases
     );
   });
+
+g.test('abstract_float')
+  .specURL('https://www.w3.org/TR/WGSL/#value-constructor-builtin-function')
+  .desc(
+    `
+f32(e), where e is an AbstractFloat
+
+Correctly rounded to f32
+`
+  )
+  .params(u =>
+    u
+      .combine('inputSource', onlyConstInputSource)
+      .combine('vectorize', [undefined, 2, 3, 4] as const)
+  )
+  .fn(async t => {
+    const cases = await d.get('abstract_float');
+    await run(
+      t,
+      vectorizeToExpression(t.params.vectorize),
+      [Type.abstractFloat],
+      Type.f32,
+      t.params,
+      cases
+    );
+  });
+
+g.test('abstract_float_mat')
+  .specURL('https://www.w3.org/TR/WGSL/#matrix-builtin-functions')
+  .desc(`AbstractFloat matrix to f32 matrix tests`)
+  .params(u =>
+    u
+      .combine('inputSource', onlyConstInputSource)
+      .combine('cols', [2, 3, 4] as const)
+      .combine('rows', [2, 3, 4] as const)
+  )
+  .fn(async t => {
+    const cols = t.params.cols;
+    const rows = t.params.rows;
+    const cases = await d.get(`abstract_float_mat${cols}x${rows}`);
+    await run(
+      t,
+      matrixExperession(cols, rows),
+      [Type.mat(cols, rows, Type.abstractFloat)],
+      Type.mat(cols, rows, Type.f32),
+      t.params,
+      cases
+    );
+  });