diff --git a/.github/workflows/Documentation.yml b/.github/workflows/Documentation.yml
index d26b9dd03..1fce3a0cb 100644
--- a/.github/workflows/Documentation.yml
+++ b/.github/workflows/Documentation.yml
@@ -11,7 +11,7 @@ jobs:
   build:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
       - uses: julia-actions/setup-julia@latest
         with:
           version: '1'
@@ -23,6 +23,6 @@ jobs:
           DOCUMENTER_KEY: ${{ secrets.DOCUMENTER_KEY }} # For authentication with SSH deploy key
         run: julia --project=docs/ --code-coverage=user docs/make.jl
       - uses: julia-actions/julia-processcoverage@v1
-      - uses: codecov/codecov-action@v4
+      - uses: codecov/codecov-action@v5
         with:
           file: lcov.info
diff --git a/.github/workflows/Downgrade.yml b/.github/workflows/Downgrade.yml
new file mode 100644
index 000000000..64466aab5
--- /dev/null
+++ b/.github/workflows/Downgrade.yml
@@ -0,0 +1,36 @@
+name: Downgrade
+on:
+  pull_request:
+    branches:
+      - master
+    paths-ignore:
+      - 'docs/**'
+      - 'benchmark/**'
+  push:
+    branches:
+      - master
+    paths-ignore:
+      - 'docs/**'
+      - 'benchmark/**'
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        downgrade_mode: ['alldeps']
+        julia-version: ['1.10']
+        group: ['Core']
+    steps:
+      - uses: actions/checkout@v5
+      - uses: julia-actions/setup-julia@v2
+        with:
+          version: ${{ matrix.julia-version }}
+      - uses: julia-actions/julia-downgrade-compat@v2
+        with:
+          skip: Pkg,TOML
+      - uses: julia-actions/julia-buildpkg@v1
+      - uses: julia-actions/julia-runtest@v1
+        with:
+          ALLOW_RERESOLVE: false
+        env:
+          GROUP: ${{ matrix.group }}
diff --git a/.github/workflows/Downstream.yml b/.github/workflows/Downstream.yml
index 34b8e1bca..dc09c0945 100644
--- a/.github/workflows/Downstream.yml
+++ b/.github/workflows/Downstream.yml
@@ -32,14 +32,14 @@ jobs:
           - {user: SciML, repo: DelayDiffEq.jl, group: Interface}
 
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
       - uses: julia-actions/setup-julia@v1
         with:
           version: ${{ matrix.julia-version }}
           arch: x64
       - uses: julia-actions/julia-buildpkg@latest
       - name: Clone Downstream
-        uses: actions/checkout@v4
+        uses: actions/checkout@v5
         with:
           repository: ${{ matrix.package.user }}/${{ matrix.package.repo }}
           path: downstream
@@ -61,6 +61,6 @@ jobs:
             exit(0)  # Exit immediately, as a success
           end
       - uses: julia-actions/julia-processcoverage@v1
-      - uses: codecov/codecov-action@v4
+      - uses: codecov/codecov-action@v5
         with:
           file: lcov.info
diff --git a/.github/workflows/Invalidations.yml b/.github/workflows/Invalidations.yml
index 28b9ce2fa..7f084d239 100644
--- a/.github/workflows/Invalidations.yml
+++ b/.github/workflows/Invalidations.yml
@@ -19,12 +19,12 @@ jobs:
     - uses: julia-actions/setup-julia@v1
       with:
         version: '1'
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@v5
     - uses: julia-actions/julia-buildpkg@v1
     - uses: julia-actions/julia-invalidations@v1
       id: invs_pr
 
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@v5
       with:
         ref: ${{ github.event.repository.default_branch }}
     - uses: julia-actions/julia-buildpkg@v1
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index d9333ffbf..8d933f0f0 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -15,8 +15,12 @@ jobs:
           - Core
         version:
           - '1'
+        os:
+          - ubuntu-latest
+          - macos-latest
+          - windows-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
       - uses: julia-actions/setup-julia@v1
         with:
           version: ${{ matrix.version }}
@@ -35,6 +39,6 @@ jobs:
         env:
           GROUP: ${{ matrix.group }}
       - uses: julia-actions/julia-processcoverage@v1
-      - uses: codecov/codecov-action@v4
+      - uses: codecov/codecov-action@v5
         with:
-          file: lcov.info
\ No newline at end of file
+          file: lcov.info
diff --git a/Project.toml b/Project.toml
index 3dfc4a97d..df62ce433 100644
--- a/Project.toml
+++ b/Project.toml
@@ -1,6 +1,6 @@
 name = "ArrayInterface"
 uuid = "4fba245c-0d91-5ea0-9b3e-6abc04ee57a9"
-version = "7.17.0"
+version = "7.22.0"
 
 [deps]
 Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
@@ -14,6 +14,7 @@ CUDSS = "45b445bb-4962-46a0-9369-b4df9d0f772e"
 ChainRules = "082447d4-558c-5d27-93f4-14fc19e9eca2"
 ChainRulesCore = "d360d2e6-b24c-11e9-a2a3-2a2ae2dbcce4"
 GPUArraysCore = "46192b85-c4d5-4398-a991-12ede77f4527"
+Metal = "dde4c033-4e86-420c-a63e-0dd931031962"
 ReverseDiff = "37e2e3b7-166d-5795-8a7a-e32c996b4267"
 SparseArrays = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
 StaticArraysCore = "1e83bf80-4336-4d27-bf5d-d5a4f845583c"
@@ -23,10 +24,11 @@ Tracker = "9f7883ad-71c0-57eb-9f7f-b5c9e6d3789c"
 ArrayInterfaceBandedMatricesExt = "BandedMatrices"
 ArrayInterfaceBlockBandedMatricesExt = "BlockBandedMatrices"
 ArrayInterfaceCUDAExt = "CUDA"
-ArrayInterfaceCUDSSExt = "CUDSS"
-ArrayInterfaceChainRulesExt = "ChainRules"
+ArrayInterfaceCUDSSExt = ["CUDSS", "CUDA"]
 ArrayInterfaceChainRulesCoreExt = "ChainRulesCore"
+ArrayInterfaceChainRulesExt = "ChainRules"
 ArrayInterfaceGPUArraysCoreExt = "GPUArraysCore"
+ArrayInterfaceMetalExt = "Metal"
 ArrayInterfaceReverseDiffExt = "ReverseDiff"
 ArrayInterfaceSparseArraysExt = "SparseArrays"
 ArrayInterfaceStaticArraysCoreExt = "StaticArraysCore"
@@ -37,12 +39,13 @@ Adapt = "4"
 BandedMatrices = "1"
 BlockBandedMatrices = "0.13"
 CUDA = "5"
-CUDSS = "0.2, 0.3"
+CUDSS = "0.6.1"
 ChainRules = "1"
 ChainRulesCore = "1"
 ChainRulesTestUtils = "1"
 GPUArraysCore = "0.1, 0.2"
 LinearAlgebra = "1.10"
+Metal = "1"
 ReverseDiff = "1"
 SparseArrays = "1.10"
 StaticArraysCore = "1"
@@ -59,6 +62,8 @@ ChainRulesTestUtils = "cdddcdb0-9152-4a09-a978-84456f9df70a"
 ComponentArrays = "b0b7db55-cfe3-40fc-9ded-d10e2dbeff66"
 FillArrays = "1a297f60-69ca-5386-bcde-b61e274b549b"
 GPUArraysCore = "46192b85-c4d5-4398-a991-12ede77f4527"
+JuliaFormatter = "98e50ef6-434e-11e9-1051-2b60c6c9e899"
+Metal = "dde4c033-4e86-420c-a63e-0dd931031962"
 Pkg = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 ReverseDiff = "37e2e3b7-166d-5795-8a7a-e32c996b4267"
diff --git a/ext/ArrayInterfaceBandedMatricesExt.jl b/ext/ArrayInterfaceBandedMatricesExt.jl
index fcd0fb7b0..5eaeb0cdf 100644
--- a/ext/ArrayInterfaceBandedMatricesExt.jl
+++ b/ext/ArrayInterfaceBandedMatricesExt.jl
@@ -46,7 +46,7 @@ function _bandsize(bandind, rowsize, colsize)
     end
 end
 
-function BandedMatrixIndex(rowsize, colsize, lowerbandwidth, upperbandwidth, isrow)
+function ArrayInterface.BandedMatrixIndex(rowsize, colsize, lowerbandwidth, upperbandwidth, isrow)
     upperbandwidth > -lowerbandwidth || throw(ErrorException("Invalid Bandwidths"))
     bandinds = upperbandwidth:-1:(-lowerbandwidth)
     bandsizes = [_bandsize(band, rowsize, colsize) for band in bandinds]
diff --git a/ext/ArrayInterfaceCUDSSExt.jl b/ext/ArrayInterfaceCUDSSExt.jl
index 01fb23953..e6a4908d2 100644
--- a/ext/ArrayInterfaceCUDSSExt.jl
+++ b/ext/ArrayInterfaceCUDSSExt.jl
@@ -2,14 +2,18 @@ module ArrayInterfaceCUDSSExt
 
 using ArrayInterface
 using CUDSS
+using CUDA
 
 function ArrayInterface.lu_instance(A::CUDSS.CuSparseMatrixCSR)
     ArrayInterface.LinearAlgebra.checksquare(A)
-    fact = CudssSolver(A, "G", 'F')
     T = eltype(A)
-    n = size(A,1)
-    x = CudssMatrix(T, n)
-    b = CudssMatrix(T, n)
+    n = size(A, 1)
+
+    # Use standard CUDA types (CuVector) instead of deprecated CudssMatrix
+    x = CUDA.CuVector{T}(undef, n)
+    b = CUDA.CuVector{T}(undef, n)
+
+    fact = CudssSolver(A, "G", 'F')
     cudss("analysis", fact, x, b)
     fact
 end
diff --git a/ext/ArrayInterfaceMetalExt.jl b/ext/ArrayInterfaceMetalExt.jl
new file mode 100644
index 000000000..c27f67f24
--- /dev/null
+++ b/ext/ArrayInterfaceMetalExt.jl
@@ -0,0 +1,15 @@
+module ArrayInterfaceMetalExt
+
+using ArrayInterface
+using Metal
+using LinearAlgebra
+
+function ArrayInterface.lu_instance(A::MtlMatrix{T}) where {T}
+    ipiv = MtlVector{Int32}(undef, 0)
+    info = zero(Int)
+    return LinearAlgebra.LU(similar(A, 0, 0), ipiv, info)
+end
+
+ArrayInterface.device(::Type{<:Metal.MtlArray}) = ArrayInterface.GPU()
+
+end # module
\ No newline at end of file
diff --git a/ext/ArrayInterfaceReverseDiffExt.jl b/ext/ArrayInterfaceReverseDiffExt.jl
index 37176824e..3a000def3 100644
--- a/ext/ArrayInterfaceReverseDiffExt.jl
+++ b/ext/ArrayInterfaceReverseDiffExt.jl
@@ -8,11 +8,8 @@ ArrayInterface.ismutable(T::Type{<:ReverseDiff.TrackedReal}) = false
 ArrayInterface.can_setindex(::Type{<:ReverseDiff.TrackedArray}) = false
 ArrayInterface.fast_scalar_indexing(::Type{<:ReverseDiff.TrackedArray}) = false
 function ArrayInterface.aos_to_soa(x::AbstractArray{<:ReverseDiff.TrackedReal, N}) where {N}
-    if length(x) > 1
-        return reshape(reduce(vcat, x), size(x))
-    else
-        return reduce(vcat,[x[1], x[1]])[1:1]
-    end
+    y = length(x) > 1 ? reduce(vcat, x) : reduce(vcat, [x[1], x[1]])[1:1]
+    return reshape(y, size(x))
 end
 
 function ArrayInterface.restructure(x::Array, y::ReverseDiff.TrackedArray)
diff --git a/ext/ArrayInterfaceSparseArraysExt.jl b/ext/ArrayInterfaceSparseArraysExt.jl
index 61fff69dc..9f0e06b43 100644
--- a/ext/ArrayInterfaceSparseArraysExt.jl
+++ b/ext/ArrayInterfaceSparseArraysExt.jl
@@ -14,25 +14,25 @@ function findstructralnz(x::SparseMatrixCSC)
     (rowind, colind)
 end
 
-function bunchkaufman_instance(A::SparseMatrixCSC)
-    bunchkaufman(sparse(similar(A, 1, 1)), check = false)
+function bunchkaufman_instance(A::SparseMatrixCSC{Tv, Ti}) where {Tv, Ti}
+    bunchkaufman(SparseMatrixCSC{Tv, Ti}(similar(A, 1, 1)), check = false)
 end
 
-function cholesky_instance(A::Union{SparseMatrixCSC,Symmetric{<:Number,<:SparseMatrixCSC}}, pivot = DEFAULT_CHOLESKY_PIVOT)
-    cholesky(sparse(similar(A, 1, 1)), check = false)
+function cholesky_instance(A::Union{SparseMatrixCSC{Tv, Ti},Symmetric{<:Number,<:SparseMatrixCSC{Tv, Ti}}}, pivot = DEFAULT_CHOLESKY_PIVOT) where {Tv, Ti}
+    cholesky(SparseMatrixCSC{Tv, Ti}(similar(A, 1, 1)), check = false)
 end
 
-function ldlt_instance(A::SparseMatrixCSC)
-    ldlt(sparse(similar(A, 1, 1)), check=false)
+function ldlt_instance(A::SparseMatrixCSC{Tv, Ti}) where {Tv, Ti}
+    ldlt(SparseMatrixCSC{Tv, Ti}(similar(A, 1, 1)), check=false)
 end
 
 # Could be optimized but this should work for any real case.
-function lu_instance(jac_prototype::SparseMatrixCSC, pivot = DEFAULT_CHOLESKY_PIVOT)
-    lu(sparse(rand(1,1)))
+function lu_instance(jac_prototype::SparseMatrixCSC{Tv, Ti}, pivot = DEFAULT_CHOLESKY_PIVOT) where {Tv, Ti}
+    lu(SparseMatrixCSC{Tv, Ti}(rand(1,1)))
 end
 
-function qr_instance(jac_prototype::SparseMatrixCSC, pivot = DEFAULT_CHOLESKY_PIVOT)
-    qr(sparse(rand(1,1)))
+function qr_instance(jac_prototype::SparseMatrixCSC{Tv, Ti}, pivot = DEFAULT_CHOLESKY_PIVOT) where {Tv, Ti}
+    qr(SparseMatrixCSC{Tv, Ti}(rand(1,1)))
 end
 
 end
diff --git a/test/ad.jl b/test/ad.jl
index bc3c3dd80..3c61873e6 100644
--- a/test/ad.jl
+++ b/test/ad.jl
@@ -1,6 +1,9 @@
 using ArrayInterface, ReverseDiff, Tracker, Test
 x = ReverseDiff.track([4.0])
 @test ArrayInterface.aos_to_soa(x) isa ReverseDiff.TrackedArray
+x = reshape([ReverseDiff.track(rand(1, 1, 1))[1]], 1, 1, 1)
+@test ArrayInterface.aos_to_soa(x) isa ReverseDiff.TrackedArray
+@test ndims(ArrayInterface.aos_to_soa(x)) == 3
 x = reduce(vcat, ReverseDiff.track([4.0,4.0]))
 @test ArrayInterface.aos_to_soa(x) isa ReverseDiff.TrackedArray
 x = [ReverseDiff.track([4.0])[1]]
diff --git a/test/gpu/Project.toml b/test/gpu/Project.toml
index 0853fd7bb..1ef4f1022 100644
--- a/test/gpu/Project.toml
+++ b/test/gpu/Project.toml
@@ -1,2 +1,4 @@
 [deps]
-CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
\ No newline at end of file
+CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
+CUDSS = "45b445bb-4962-46a0-9369-b4df9d0f772e"
+Metal = "dde4c033-4e86-420c-a63e-0dd931031962"
diff --git a/test/gpu/cuda.jl b/test/gpu/cuda.jl
index f57bdafa7..2651a6381 100644
--- a/test/gpu/cuda.jl
+++ b/test/gpu/cuda.jl
@@ -1,7 +1,20 @@
-using CUDA
+using CUDSS, CUDA, SparseArrays, LinearAlgebra
+using CUDA.CUSPARSE
 using ArrayInterface
 
 using Test
 
+A_cpu = Float32[1 0; 0 1]
+A_dense = CuMatrix(A_cpu)
+A_sparse = CuSparseMatrixCSR(sparse(A_cpu))
+
 # Test whether lu_instance throws an error when invoked with an gpu array
-@test !isa(try ArrayInterface.lu_instance(CUDA.CuArray([1.f0 1.f0; 1.f0 1.f0])) catch ex ex end, Exception) 
+lu_inst_dense = ArrayInterface.lu_instance(A_dense)
+lu_inst_sparse = ArrayInterface.lu_instance(A_sparse)
+
+# test that lu! is valid when using the inst as scratch
+lu_sparse = lu!(lu_inst_sparse, A_sparse)
+
+#test that the resulting lu works
+b = CuVector([1f0, 1f0])
+@test CUDA.@allowscalar lu_sparse \ b == [1, 1]
diff --git a/test/gpu/metal.jl b/test/gpu/metal.jl
new file mode 100644
index 000000000..c819a05ee
--- /dev/null
+++ b/test/gpu/metal.jl
@@ -0,0 +1,8 @@
+using Metal
+using ArrayInterface
+using LinearAlgebra
+
+using Test
+
+# Test that lu_instance works with Metal.jl gpu arrays
+@test isa(ArrayInterface.lu_instance(MtlArray([1.f0 1.f0; 1.f0 1.f0])), LU) 
\ No newline at end of file
diff --git a/test/runtests.jl b/test/runtests.jl
index 8a5d7b363..cd47c19fb 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -21,5 +21,8 @@ end
     if GROUP == "GPU"
         activate_gpu_env()
         @time @safetestset "CUDA" begin include("gpu/cuda.jl") end
+        if Sys.isapple()
+            @time @safetestset "Metal" begin include("gpu/metal.jl") end
+        end
     end
 end