diff --git a/src/libraries/System.Numerics.Tensors/ref/System.Numerics.Tensors.netcore.cs b/src/libraries/System.Numerics.Tensors/ref/System.Numerics.Tensors.netcore.cs
index 40fcdafe65436..e7945587afef4 100644
--- a/src/libraries/System.Numerics.Tensors/ref/System.Numerics.Tensors.netcore.cs
+++ b/src/libraries/System.Numerics.Tensors/ref/System.Numerics.Tensors.netcore.cs
@@ -89,12 +89,18 @@ public static partial class TensorPrimitives
         public static T Max<T>(System.ReadOnlySpan<T> x) where T : System.Numerics.INumber<T> { throw null; }
         public static void Max<T>(System.ReadOnlySpan<T> x, System.ReadOnlySpan<T> y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
         public static void Max<T>(System.ReadOnlySpan<T> x, T y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
+        public static T MaxNumber<T>(System.ReadOnlySpan<T> x) where T : System.Numerics.INumber<T> { throw null; }
+        public static void MaxNumber<T>(System.ReadOnlySpan<T> x, System.ReadOnlySpan<T> y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
+        public static void MaxNumber<T>(System.ReadOnlySpan<T> x, T y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
         public static T MinMagnitude<T>(System.ReadOnlySpan<T> x) where T : System.Numerics.INumberBase<T> { throw null; }
         public static void MinMagnitude<T>(System.ReadOnlySpan<T> x, System.ReadOnlySpan<T> y, System.Span<T> destination) where T : System.Numerics.INumberBase<T> { }
         public static void MinMagnitude<T>(System.ReadOnlySpan<T> x, T y, System.Span<T> destination) where T : System.Numerics.INumberBase<T> { }
         public static T Min<T>(System.ReadOnlySpan<T> x) where T : System.Numerics.INumber<T> { throw null; }
         public static void Min<T>(System.ReadOnlySpan<T> x, System.ReadOnlySpan<T> y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
         public static void Min<T>(System.ReadOnlySpan<T> x, T y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
+        public static T MinNumber<T>(System.ReadOnlySpan<T> x) where T : System.Numerics.INumber<T> { throw null; }
+        public static void MinNumber<T>(System.ReadOnlySpan<T> x, System.ReadOnlySpan<T> y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
+        public static void MinNumber<T>(System.ReadOnlySpan<T> x, T y, System.Span<T> destination) where T : System.Numerics.INumber<T> { }
         public static void MultiplyAdd<T>(System.ReadOnlySpan<T> x, System.ReadOnlySpan<T> y, System.ReadOnlySpan<T> addend, System.Span<T> destination) where T : System.Numerics.IAdditionOperators<T, T, T>, System.Numerics.IMultiplyOperators<T, T, T> { }
         public static void MultiplyAdd<T>(System.ReadOnlySpan<T> x, System.ReadOnlySpan<T> y, T addend, System.Span<T> destination) where T : System.Numerics.IAdditionOperators<T, T, T>, System.Numerics.IMultiplyOperators<T, T, T> { }
         public static void MultiplyAdd<T>(System.ReadOnlySpan<T> x, T y, System.ReadOnlySpan<T> addend, System.Span<T> destination) where T : System.Numerics.IAdditionOperators<T, T, T>, System.Numerics.IMultiplyOperators<T, T, T> { }
diff --git a/src/libraries/System.Numerics.Tensors/src/System.Numerics.Tensors.csproj b/src/libraries/System.Numerics.Tensors/src/System.Numerics.Tensors.csproj
index f5bb6c851dfb0..bc0e5be9bed57 100644
--- a/src/libraries/System.Numerics.Tensors/src/System.Numerics.Tensors.csproj
+++ b/src/libraries/System.Numerics.Tensors/src/System.Numerics.Tensors.csproj
@@ -1,4 +1,4 @@
-<Project Sdk="Microsoft.NET.Sdk">
+<Project Sdk="Microsoft.NET.Sdk">
   
   <PropertyGroup>
     <TargetFrameworks>$(NetCoreAppCurrent);$(NetCoreAppPrevious);$(NetCoreAppMinimum);netstandard2.0;$(NetFrameworkMinimum)</TargetFrameworks>
@@ -82,8 +82,10 @@
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.LogP1.cs" />
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.Max.cs" />
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.MaxMagnitude.cs" />
+    <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.MaxNumber.cs" />
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.Min.cs" />
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.MinMagnitude.cs" />
+    <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.MinNumber.cs" />
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.Multiply.cs" />
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.MultiplyAdd.cs" />
     <Compile Include="System\Numerics\Tensors\netcore\TensorPrimitives.MultiplyAddEstimate.cs" />
diff --git a/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Max.cs b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Max.cs
index 0e0566f2d935a..2e760e30fb7c6 100644
--- a/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Max.cs
+++ b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Max.cs
@@ -96,35 +96,21 @@ public static T Invoke(T x, T y)
             [MethodImpl(MethodImplOptions.AggressiveInlining)]
             public static Vector128<T> Invoke(Vector128<T> x, Vector128<T> y)
             {
-                if (AdvSimd.IsSupported)
-                {
-                    if (typeof(T) == typeof(byte)) return AdvSimd.Max(x.AsByte(), y.AsByte()).As<byte, T>();
-                    if (typeof(T) == typeof(sbyte)) return AdvSimd.Max(x.AsSByte(), y.AsSByte()).As<sbyte, T>();
-                    if (typeof(T) == typeof(short)) return AdvSimd.Max(x.AsInt16(), y.AsInt16()).As<short, T>();
-                    if (typeof(T) == typeof(ushort)) return AdvSimd.Max(x.AsUInt16(), y.AsUInt16()).As<ushort, T>();
-                    if (typeof(T) == typeof(int)) return AdvSimd.Max(x.AsInt32(), y.AsInt32()).As<int, T>();
-                    if (typeof(T) == typeof(uint)) return AdvSimd.Max(x.AsUInt32(), y.AsUInt32()).As<uint, T>();
-                    if (typeof(T) == typeof(float)) return AdvSimd.Max(x.AsSingle(), y.AsSingle()).As<float, T>();
-                }
-
-                if (AdvSimd.Arm64.IsSupported)
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
                 {
-                    if (typeof(T) == typeof(double)) return AdvSimd.Arm64.Max(x.AsDouble(), y.AsDouble()).As<double, T>();
-                }
+                    if (AdvSimd.IsSupported && typeof(T) == typeof(float))
+                    {
+                        return AdvSimd.Max(x.AsSingle(), y.AsSingle()).As<float, T>();
+                    }
 
-                if (typeof(T) == typeof(float))
-                {
-                    return
-                        Vector128.ConditionalSelect(Vector128.Equals(x, y),
-                            Vector128.ConditionalSelect(IsNegative(x.AsSingle()).As<float, T>(), y, x),
-                            Vector128.Max(x, y));
-                }
+                    if (AdvSimd.Arm64.IsSupported && typeof(T) == typeof(double))
+                    {
+                        return AdvSimd.Arm64.Max(x.AsDouble(), y.AsDouble()).As<double, T>();
+                    }
 
-                if (typeof(T) == typeof(double))
-                {
                     return
                         Vector128.ConditionalSelect(Vector128.Equals(x, y),
-                            Vector128.ConditionalSelect(IsNegative(x.AsDouble()).As<double, T>(), y, x),
+                            Vector128.ConditionalSelect(IsNegative(x), y, x),
                             Vector128.Max(x, y));
                 }
 
@@ -134,19 +120,11 @@ public static Vector128<T> Invoke(Vector128<T> x, Vector128<T> y)
             [MethodImpl(MethodImplOptions.AggressiveInlining)]
             public static Vector256<T> Invoke(Vector256<T> x, Vector256<T> y)
             {
-                if (typeof(T) == typeof(float))
-                {
-                    return
-                        Vector256.ConditionalSelect(Vector256.Equals(x, y),
-                            Vector256.ConditionalSelect(IsNegative(x.AsSingle()).As<float, T>(), y, x),
-                            Vector256.Max(x, y));
-                }
-
-                if (typeof(T) == typeof(double))
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
                 {
                     return
                         Vector256.ConditionalSelect(Vector256.Equals(x, y),
-                            Vector256.ConditionalSelect(IsNegative(x.AsDouble()).As<double, T>(), y, x),
+                            Vector256.ConditionalSelect(IsNegative(x), y, x),
                             Vector256.Max(x, y));
                 }
 
@@ -156,19 +134,11 @@ public static Vector256<T> Invoke(Vector256<T> x, Vector256<T> y)
             [MethodImpl(MethodImplOptions.AggressiveInlining)]
             public static Vector512<T> Invoke(Vector512<T> x, Vector512<T> y)
             {
-                if (typeof(T) == typeof(float))
-                {
-                    return
-                        Vector512.ConditionalSelect(Vector512.Equals(x, y),
-                            Vector512.ConditionalSelect(IsNegative(x.AsSingle()).As<float, T>(), y, x),
-                            Vector512.Max(x, y));
-                }
-
-                if (typeof(T) == typeof(double))
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
                 {
                     return
                         Vector512.ConditionalSelect(Vector512.Equals(x, y),
-                            Vector512.ConditionalSelect(IsNegative(x.AsDouble()).As<double, T>(), y, x),
+                            Vector512.ConditionalSelect(IsNegative(x), y, x),
                             Vector512.Max(x, y));
                 }
 
@@ -192,24 +162,18 @@ public static Vector512<T> Invoke(Vector512<T> x, Vector512<T> y)
             [MethodImpl(MethodImplOptions.AggressiveInlining)]
             public static Vector128<T> Invoke(Vector128<T> x, Vector128<T> y)
             {
-                if (AdvSimd.IsSupported)
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
                 {
-                    if (typeof(T) == typeof(byte)) return AdvSimd.Max(x.AsByte(), y.AsByte()).As<byte, T>();
-                    if (typeof(T) == typeof(sbyte)) return AdvSimd.Max(x.AsSByte(), y.AsSByte()).As<sbyte, T>();
-                    if (typeof(T) == typeof(ushort)) return AdvSimd.Max(x.AsUInt16(), y.AsUInt16()).As<ushort, T>();
-                    if (typeof(T) == typeof(short)) return AdvSimd.Max(x.AsInt16(), y.AsInt16()).As<short, T>();
-                    if (typeof(T) == typeof(uint)) return AdvSimd.Max(x.AsUInt32(), y.AsUInt32()).As<uint, T>();
-                    if (typeof(T) == typeof(int)) return AdvSimd.Max(x.AsInt32(), y.AsInt32()).As<int, T>();
-                    if (typeof(T) == typeof(float)) return AdvSimd.Max(x.AsSingle(), y.AsSingle()).As<float, T>();
-                }
+                    if (AdvSimd.IsSupported && typeof(T) == typeof(float))
+                    {
+                        return AdvSimd.Max(x.AsSingle(), y.AsSingle()).As<float, T>();
+                    }
 
-                if (AdvSimd.Arm64.IsSupported)
-                {
-                    if (typeof(T) == typeof(double)) return AdvSimd.Arm64.Max(x.AsDouble(), y.AsDouble()).As<double, T>();
-                }
+                    if (AdvSimd.Arm64.IsSupported && typeof(T) == typeof(double))
+                    {
+                        return AdvSimd.Arm64.Max(x.AsDouble(), y.AsDouble()).As<double, T>();
+                    }
 
-                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
-                {
                     return
                         Vector128.ConditionalSelect(Vector128.Equals(x, x),
                             Vector128.ConditionalSelect(Vector128.Equals(y, y),
diff --git a/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.MaxNumber.cs b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.MaxNumber.cs
new file mode 100644
index 0000000000000..ef4b39a77f8cb
--- /dev/null
+++ b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.MaxNumber.cs
@@ -0,0 +1,142 @@
+// Licensed to the .NET Foundation under one or more agreements.
+// The .NET Foundation licenses this file to you under the MIT license.
+
+using System.Runtime.CompilerServices;
+using System.Runtime.InteropServices;
+using System.Runtime.Intrinsics;
+using System.Runtime.Intrinsics.Arm;
+
+namespace System.Numerics.Tensors
+{
+    public static partial class TensorPrimitives
+    {
+        /// <summary>Searches for the largest number in the specified tensor.</summary>
+        /// <param name="x">The tensor, represented as a span.</param>
+        /// <returns>The maximum element in <paramref name="x"/>.</returns>
+        /// <exception cref="ArgumentException">Length of <paramref name="x" /> must be greater than zero.</exception>
+        /// <remarks>
+        /// <para>
+        /// The determination of the maximum element matches the IEEE 754:2019 `maximumNumber` function. Positive 0 is considered greater than negative 0.
+        /// </para>
+        /// <para>
+        /// This method may call into the underlying C runtime or employ instructions specific to the current architecture. Exact results may differ between different
+        /// operating systems or architectures.
+        /// </para>
+        /// </remarks>
+        public static T MaxNumber<T>(ReadOnlySpan<T> x)
+            where T : INumber<T> =>
+            MinMaxCore<T, MaxNumberOperator<T>>(x);
+
+        /// <summary>Computes the element-wise maximum of the numbers in the specified tensors.</summary>
+        /// <param name="x">The first tensor, represented as a span.</param>
+        /// <param name="y">The second tensor, represented as a span.</param>
+        /// <param name="destination">The destination tensor, represented as a span.</param>
+        /// <exception cref="ArgumentException">Length of <paramref name="x" /> must be same as length of <paramref name="y" />.</exception>
+        /// <exception cref="ArgumentException">Destination is too short.</exception>
+        /// <exception cref="ArgumentException"><paramref name="x"/> and <paramref name="destination"/> reference overlapping memory locations and do not begin at the same location.</exception>
+        /// <exception cref="ArgumentException"><paramref name="y"/> and <paramref name="destination"/> reference overlapping memory locations and do not begin at the same location.</exception>
+        /// <remarks>
+        /// <para>
+        /// This method effectively computes <c><paramref name="destination" />[i] = <typeparamref name="T"/>.MaxNumber(<paramref name="x" />[i], <paramref name="y" />[i])</c>.
+        /// </para>
+        /// <para>
+        /// The determination of the maximum element matches the IEEE 754:2019 `maximumNumber` function. If either value is <see cref="IFloatingPointIeee754{TSelf}.NaN"/>
+        /// the other is returned. Positive 0 is considered greater than negative 0.
+        /// </para>
+        /// <para>
+        /// This method may call into the underlying C runtime or employ instructions specific to the current architecture. Exact results may differ between different
+        /// operating systems or architectures.
+        /// </para>
+        /// </remarks>
+        public static void MaxNumber<T>(ReadOnlySpan<T> x, ReadOnlySpan<T> y, Span<T> destination)
+            where T : INumber<T> =>
+            InvokeSpanSpanIntoSpan<T, MaxNumberOperator<T>>(x, y, destination);
+
+        /// <summary>Computes the element-wise maximum of the numbers in the specified tensors.</summary>
+        /// <param name="x">The first tensor, represented as a span.</param>
+        /// <param name="y">The second tensor, represented as a scalar.</param>
+        /// <param name="destination">The destination tensor, represented as a span.</param>
+        /// <exception cref="ArgumentException">Destination is too short.</exception>
+        /// <exception cref="ArgumentException"><paramref name="x"/> and <paramref name="destination"/> reference overlapping memory locations and do not begin at the same location.</exception>
+        /// <remarks>
+        /// <para>
+        /// This method effectively computes <c><paramref name="destination" />[i] = <typeparamref name="T"/>.MaxNumber(<paramref name="x" />[i], <paramref name="y" />)</c>.
+        /// </para>
+        /// <para>
+        /// The determination of the maximum element matches the IEEE 754:2019 `maximumNumber` function. If either value is <see cref="IFloatingPointIeee754{TSelf}.NaN"/>
+        /// the other is returned. Positive 0 is considered greater than negative 0.
+        /// </para>
+        /// <para>
+        /// This method may call into the underlying C runtime or employ instructions specific to the current architecture. Exact results may differ between different
+        /// operating systems or architectures.
+        /// </para>
+        /// </remarks>
+        public static void MaxNumber<T>(ReadOnlySpan<T> x, T y, Span<T> destination)
+            where T : INumber<T> =>
+            InvokeSpanScalarIntoSpan<T, MaxNumberOperator<T>>(x, y, destination);
+
+        /// <summary>T.MaxNumber(x, y)</summary>
+        internal readonly struct MaxNumberOperator<T> : IAggregationOperator<T> where T : INumber<T>
+        {
+            public static bool Vectorizable => true;
+
+            public static T Invoke(T x, T y) => T.MaxNumber(x, y);
+
+            [MethodImpl(MethodImplOptions.AggressiveInlining)]
+            public static Vector128<T> Invoke(Vector128<T> x, Vector128<T> y)
+            {
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
+                {
+                    if (AdvSimd.IsSupported && typeof(T) == typeof(float))
+                    {
+                        return AdvSimd.MaxNumber(x.AsSingle(), y.AsSingle()).As<float, T>();
+                    }
+
+                    if (AdvSimd.Arm64.IsSupported && typeof(T) == typeof(double))
+                    {
+                        return AdvSimd.Arm64.MaxNumber(x.AsDouble(), y.AsDouble()).As<double, T>();
+                    }
+
+                    return
+                        Vector128.ConditionalSelect(Vector128.Equals(x, y),
+                            Vector128.ConditionalSelect(IsNegative(y), x, y),
+                            Vector128.ConditionalSelect(Vector128.Equals(y, y), Vector128.Max(x, y), x));
+                }
+
+                return Vector128.Max(x, y);
+            }
+
+            [MethodImpl(MethodImplOptions.AggressiveInlining)]
+            public static Vector256<T> Invoke(Vector256<T> x, Vector256<T> y)
+            {
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
+                {
+                    return
+                        Vector256.ConditionalSelect(Vector256.Equals(x, y),
+                            Vector256.ConditionalSelect(IsNegative(y), x, y),
+                            Vector256.ConditionalSelect(Vector256.Equals(y, y), Vector256.Max(x, y), x));
+                }
+
+                return Vector256.Max(x, y);
+            }
+
+            [MethodImpl(MethodImplOptions.AggressiveInlining)]
+            public static Vector512<T> Invoke(Vector512<T> x, Vector512<T> y)
+            {
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
+                {
+                    return
+                        Vector512.ConditionalSelect(Vector512.Equals(x, y),
+                            Vector512.ConditionalSelect(IsNegative(y), x, y),
+                            Vector512.ConditionalSelect(Vector512.Equals(y, y), Vector512.Max(x, y), x));
+                }
+
+                return Vector512.Max(x, y);
+            }
+
+            public static T Invoke(Vector128<T> x) => HorizontalAggregate<T, MaxNumberOperator<T>>(x);
+            public static T Invoke(Vector256<T> x) => HorizontalAggregate<T, MaxNumberOperator<T>>(x);
+            public static T Invoke(Vector512<T> x) => HorizontalAggregate<T, MaxNumberOperator<T>>(x);
+        }
+    }
+}
diff --git a/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Min.cs b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Min.cs
index faea478057982..fd3931fa3d194 100644
--- a/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Min.cs
+++ b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.Min.cs
@@ -79,24 +79,18 @@ public static T Invoke(T x, T y)
             [MethodImpl(MethodImplOptions.AggressiveInlining)]
             public static Vector128<T> Invoke(Vector128<T> x, Vector128<T> y)
             {
-                if (AdvSimd.IsSupported)
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
                 {
-                    if (typeof(T) == typeof(byte)) return AdvSimd.Min(x.AsByte(), y.AsByte()).As<byte, T>();
-                    if (typeof(T) == typeof(sbyte)) return AdvSimd.Min(x.AsSByte(), y.AsSByte()).As<sbyte, T>();
-                    if (typeof(T) == typeof(short)) return AdvSimd.Min(x.AsInt16(), y.AsInt16()).As<short, T>();
-                    if (typeof(T) == typeof(ushort)) return AdvSimd.Min(x.AsUInt16(), y.AsUInt16()).As<ushort, T>();
-                    if (typeof(T) == typeof(int)) return AdvSimd.Min(x.AsInt32(), y.AsInt32()).As<int, T>();
-                    if (typeof(T) == typeof(uint)) return AdvSimd.Min(x.AsUInt32(), y.AsUInt32()).As<uint, T>();
-                    if (typeof(T) == typeof(float)) return AdvSimd.Min(x.AsSingle(), y.AsSingle()).As<float, T>();
-                }
+                    if (AdvSimd.IsSupported && typeof(T) == typeof(float))
+                    {
+                        return AdvSimd.Min(x.AsSingle(), y.AsSingle()).As<float, T>();
+                    }
 
-                if (AdvSimd.Arm64.IsSupported)
-                {
-                    if (typeof(T) == typeof(double)) return AdvSimd.Arm64.Min(x.AsDouble(), y.AsDouble()).As<double, T>();
-                }
+                    if (AdvSimd.Arm64.IsSupported && typeof(T) == typeof(double))
+                    {
+                        return AdvSimd.Arm64.Min(x.AsDouble(), y.AsDouble()).As<double, T>();
+                    }
 
-                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
-                {
                     return
                         Vector128.ConditionalSelect(Vector128.Equals(x, y),
                             Vector128.ConditionalSelect(IsNegative(y), y, x),
@@ -149,24 +143,18 @@ public static Vector512<T> Invoke(Vector512<T> x, Vector512<T> y)
             [MethodImpl(MethodImplOptions.AggressiveInlining)]
             public static Vector128<T> Invoke(Vector128<T> x, Vector128<T> y)
             {
-                if (AdvSimd.IsSupported)
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
                 {
-                    if (typeof(T) == typeof(byte)) return AdvSimd.Min(x.AsByte(), y.AsByte()).As<byte, T>();
-                    if (typeof(T) == typeof(sbyte)) return AdvSimd.Min(x.AsSByte(), y.AsSByte()).As<sbyte, T>();
-                    if (typeof(T) == typeof(short)) return AdvSimd.Min(x.AsInt16(), y.AsInt16()).As<short, T>();
-                    if (typeof(T) == typeof(ushort)) return AdvSimd.Min(x.AsUInt16(), y.AsUInt16()).As<ushort, T>();
-                    if (typeof(T) == typeof(int)) return AdvSimd.Min(x.AsInt32(), y.AsInt32()).As<int, T>();
-                    if (typeof(T) == typeof(uint)) return AdvSimd.Min(x.AsUInt32(), y.AsUInt32()).As<uint, T>();
-                    if (typeof(T) == typeof(float)) return AdvSimd.Min(x.AsSingle(), y.AsSingle()).As<float, T>();
-                }
+                    if (AdvSimd.IsSupported && typeof(T) == typeof(float))
+                    {
+                        return AdvSimd.Min(x.AsSingle(), y.AsSingle()).As<float, T>();
+                    }
 
-                if (AdvSimd.Arm64.IsSupported)
-                {
-                    if (typeof(T) == typeof(double)) return AdvSimd.Arm64.Min(x.AsDouble(), y.AsDouble()).As<double, T>();
-                }
+                    if (AdvSimd.Arm64.IsSupported && typeof(T) == typeof(double))
+                    {
+                        return AdvSimd.Arm64.Min(x.AsDouble(), y.AsDouble()).As<double, T>();
+                    }
 
-                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
-                {
                     return
                         Vector128.ConditionalSelect(Vector128.Equals(x, x),
                             Vector128.ConditionalSelect(Vector128.Equals(y, y),
diff --git a/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.MinNumber.cs b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.MinNumber.cs
new file mode 100644
index 0000000000000..ab0cb5080fab0
--- /dev/null
+++ b/src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/TensorPrimitives.MinNumber.cs
@@ -0,0 +1,142 @@
+// Licensed to the .NET Foundation under one or more agreements.
+// The .NET Foundation licenses this file to you under the MIT license.
+
+using System.Runtime.CompilerServices;
+using System.Runtime.InteropServices;
+using System.Runtime.Intrinsics;
+using System.Runtime.Intrinsics.Arm;
+
+namespace System.Numerics.Tensors
+{
+    public static partial class TensorPrimitives
+    {
+        /// <summary>Searches for the largest number in the specified tensor.</summary>
+        /// <param name="x">The tensor, represented as a span.</param>
+        /// <returns>The maximum element in <paramref name="x"/>.</returns>
+        /// <exception cref="ArgumentException">Length of <paramref name="x" /> must be greater than zero.</exception>
+        /// <remarks>
+        /// <para>
+        /// The determination of the maximum element matches the IEEE 754:2019 `maximumNumber` function. Positive 0 is considered greater than negative 0.
+        /// </para>
+        /// <para>
+        /// This method may call into the underlying C runtime or employ instructions specific to the current architecture. Exact results may differ between different
+        /// operating systems or architectures.
+        /// </para>
+        /// </remarks>
+        public static T MinNumber<T>(ReadOnlySpan<T> x)
+            where T : INumber<T> =>
+            MinMaxCore<T, MinNumberOperator<T>>(x);
+
+        /// <summary>Computes the element-wise maximum of the numbers in the specified tensors.</summary>
+        /// <param name="x">The first tensor, represented as a span.</param>
+        /// <param name="y">The second tensor, represented as a span.</param>
+        /// <param name="destination">The destination tensor, represented as a span.</param>
+        /// <exception cref="ArgumentException">Length of <paramref name="x" /> must be same as length of <paramref name="y" />.</exception>
+        /// <exception cref="ArgumentException">Destination is too short.</exception>
+        /// <exception cref="ArgumentException"><paramref name="x"/> and <paramref name="destination"/> reference overlapping memory locations and do not begin at the same location.</exception>
+        /// <exception cref="ArgumentException"><paramref name="y"/> and <paramref name="destination"/> reference overlapping memory locations and do not begin at the same location.</exception>
+        /// <remarks>
+        /// <para>
+        /// This method effectively computes <c><paramref name="destination" />[i] = <typeparamref name="T"/>.MinNumber(<paramref name="x" />[i], <paramref name="y" />[i])</c>.
+        /// </para>
+        /// <para>
+        /// The determination of the maximum element matches the IEEE 754:2019 `maximumNumber` function. If either value is <see cref="IFloatingPointIeee754{TSelf}.NaN"/>
+        /// the other is returned. Positive 0 is considered greater than negative 0.
+        /// </para>
+        /// <para>
+        /// This method may call into the underlying C runtime or employ instructions specific to the current architecture. Exact results may differ between different
+        /// operating systems or architectures.
+        /// </para>
+        /// </remarks>
+        public static void MinNumber<T>(ReadOnlySpan<T> x, ReadOnlySpan<T> y, Span<T> destination)
+            where T : INumber<T> =>
+            InvokeSpanSpanIntoSpan<T, MinNumberOperator<T>>(x, y, destination);
+
+        /// <summary>Computes the element-wise maximum of the numbers in the specified tensors.</summary>
+        /// <param name="x">The first tensor, represented as a span.</param>
+        /// <param name="y">The second tensor, represented as a scalar.</param>
+        /// <param name="destination">The destination tensor, represented as a span.</param>
+        /// <exception cref="ArgumentException">Destination is too short.</exception>
+        /// <exception cref="ArgumentException"><paramref name="x"/> and <paramref name="destination"/> reference overlapping memory locations and do not begin at the same location.</exception>
+        /// <remarks>
+        /// <para>
+        /// This method effectively computes <c><paramref name="destination" />[i] = <typeparamref name="T"/>.MinNumber(<paramref name="x" />[i], <paramref name="y" />)</c>.
+        /// </para>
+        /// <para>
+        /// The determination of the maximum element matches the IEEE 754:2019 `maximumNumber` function. If either value is <see cref="IFloatingPointIeee754{TSelf}.NaN"/>
+        /// the other is returned. Positive 0 is considered greater than negative 0.
+        /// </para>
+        /// <para>
+        /// This method may call into the underlying C runtime or employ instructions specific to the current architecture. Exact results may differ between different
+        /// operating systems or architectures.
+        /// </para>
+        /// </remarks>
+        public static void MinNumber<T>(ReadOnlySpan<T> x, T y, Span<T> destination)
+            where T : INumber<T> =>
+            InvokeSpanScalarIntoSpan<T, MinNumberOperator<T>>(x, y, destination);
+
+        /// <summary>T.MinNumber(x, y)</summary>
+        internal readonly struct MinNumberOperator<T> : IAggregationOperator<T> where T : INumber<T>
+        {
+            public static bool Vectorizable => true;
+
+            public static T Invoke(T x, T y) => T.MinNumber(x, y);
+
+            [MethodImpl(MethodImplOptions.AggressiveInlining)]
+            public static Vector128<T> Invoke(Vector128<T> x, Vector128<T> y)
+            {
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
+                {
+                    if (AdvSimd.IsSupported && typeof(T) == typeof(float))
+                    {
+                        return AdvSimd.MinNumber(x.AsSingle(), y.AsSingle()).As<float, T>();
+                    }
+
+                    if (AdvSimd.Arm64.IsSupported && typeof(T) == typeof(double))
+                    {
+                        return AdvSimd.Arm64.MinNumber(x.AsDouble(), y.AsDouble()).As<double, T>();
+                    }
+
+                    return
+                        Vector128.ConditionalSelect(Vector128.Equals(x, y),
+                            Vector128.ConditionalSelect(IsNegative(x), x, y),
+                            Vector128.ConditionalSelect(Vector128.Equals(y, y), Vector128.Min(x, y), x));
+                }
+
+                return Vector128.Min(x, y);
+            }
+
+            [MethodImpl(MethodImplOptions.AggressiveInlining)]
+            public static Vector256<T> Invoke(Vector256<T> x, Vector256<T> y)
+            {
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
+                {
+                    return
+                        Vector256.ConditionalSelect(Vector256.Equals(x, y),
+                            Vector256.ConditionalSelect(IsNegative(x), x, y),
+                            Vector256.ConditionalSelect(Vector256.Equals(y, y), Vector256.Min(x, y), x));
+                }
+
+                return Vector256.Min(x, y);
+            }
+
+            [MethodImpl(MethodImplOptions.AggressiveInlining)]
+            public static Vector512<T> Invoke(Vector512<T> x, Vector512<T> y)
+            {
+                if (typeof(T) == typeof(float) || typeof(T) == typeof(double))
+                {
+                    return
+                        Vector512.ConditionalSelect(Vector512.Equals(x, y),
+                            Vector512.ConditionalSelect(IsNegative(x), x, y),
+                            Vector512.ConditionalSelect(Vector512.Equals(y, y), Vector512.Min(x, y), x));
+                }
+
+                return Vector512.Min(x, y);
+            }
+
+            public static T Invoke(Vector128<T> x) => HorizontalAggregate<T, MinNumberOperator<T>>(x);
+            public static T Invoke(Vector256<T> x) => HorizontalAggregate<T, MinNumberOperator<T>>(x);
+            public static T Invoke(Vector512<T> x) => HorizontalAggregate<T, MinNumberOperator<T>>(x);
+        }
+    }
+}
diff --git a/src/libraries/System.Numerics.Tensors/tests/TensorPrimitives.Generic.cs b/src/libraries/System.Numerics.Tensors/tests/TensorPrimitives.Generic.cs
index 0cd20f1d647a4..0ad762f9401ed 100644
--- a/src/libraries/System.Numerics.Tensors/tests/TensorPrimitives.Generic.cs
+++ b/src/libraries/System.Numerics.Tensors/tests/TensorPrimitives.Generic.cs
@@ -511,6 +511,8 @@ public static IEnumerable<object[]> SpanSpanDestinationFunctionsToTest()
             yield return Create(TensorPrimitives.Hypot, T.Hypot);
             yield return Create(TensorPrimitives.Ieee754Remainder, T.Ieee754Remainder);
             yield return Create(TensorPrimitives.Log, T.Log);
+            yield return Create(TensorPrimitives.MaxNumber, T.MaxNumber);
+            yield return Create(TensorPrimitives.MinNumber, T.MinNumber);
             yield return Create(TensorPrimitives.Pow, T.Pow, Helpers.DetermineTolerance<T>(doubleTolerance: 1e-13, floatTolerance: 1e-5f));
 
             static object[] Create(SpanSpanDestinationDelegate tensorPrimitivesMethod, Func<T, T, T> expectedMethod, T? tolerance = null)
@@ -644,8 +646,10 @@ public static IEnumerable<object[]> SpanScalarDestinationFunctionsToTest()
             yield return Create(TensorPrimitives.Log, T.Log);
             yield return Create(TensorPrimitives.Max, T.Max);
             yield return Create(TensorPrimitives.MaxMagnitude, T.MaxMagnitude);
+            yield return Create(TensorPrimitives.MaxNumber, T.MaxNumber);
             yield return Create(TensorPrimitives.Min, T.Min);
             yield return Create(TensorPrimitives.MinMagnitude, T.MinMagnitude);
+            yield return Create(TensorPrimitives.MinNumber, T.MinNumber);
 
             static object[] Create(SpanScalarDestinationDelegate<T, T, T> tensorPrimitivesMethod, Func<T, T, T> expectedMethod, T? tolerance = null)
                 => new object[] { tensorPrimitivesMethod, expectedMethod, tolerance };
@@ -1806,8 +1810,10 @@ public static IEnumerable<object[]> SpanScalarDestinationFunctionsToTest()
             yield return Create(TensorPrimitives.BitwiseOr, (x, y) => x | y);
             yield return Create(TensorPrimitives.Max, T.Max);
             yield return Create(TensorPrimitives.MaxMagnitude, T.MaxMagnitude);
+            yield return Create(TensorPrimitives.MaxNumber, T.MaxNumber);
             yield return Create(TensorPrimitives.Min, T.Min);
             yield return Create(TensorPrimitives.MinMagnitude, T.MinMagnitude);
+            yield return Create(TensorPrimitives.MinNumber, T.MinNumber);
             yield return Create(TensorPrimitives.Xor, (x, y) => x ^ y);
 
             static object[] Create(SpanScalarDestinationDelegate<T, T, T> tensorPrimitivesMethod, Func<T, T, T> expectedMethod)