diff --git a/test/test_slices.jl b/test/test_slices.jl
new file mode 100644
index 0000000..e7f0066
--- /dev/null
+++ b/test/test_slices.jl
@@ -0,0 +1,308 @@
+module TestSlices
+
+using Test: Test
+import CTParser
+import CTBase.Exceptions
+import CTModels.OCP
+
+# for the @def macro
+import CTBase
+import CTModels
+import ExaModels
+import MadNLP
+
+include(joinpath(@__DIR__, "utils.jl"))
+
+const VERBOSE = isdefined(Main, :TestData) ? Main.TestData.VERBOSE : true
+const SHOWTIMING = isdefined(Main, :TestData) ? Main.TestData.SHOWTIMING : true
+
+function test_slices()
+    Test.@testset "Slice with end Tests" verbose=VERBOSE showtiming=SHOWTIMING begin
+
+        # ====================================================================
+        # MODEL BUILDERS - Slices in constraints
+        # ====================================================================
+        # These models test slice syntax in initial conditions: x[...](0) == x0
+        # - o1: x[1:3](0)   → literal integer range (works)
+        # - o2: x[1:end](0) → end keyword (broken - not yet supported)
+        # - o3: x[1:n](0)   → variable n (broken - not yet supported)
+
+        function get_model_o1()
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x[1:3](0) == [1, 0, 0]
+                ẋ(t) == [x₂(t), x₃(t), u(t)]
+                ∫(u(t)^2) → min
+            end
+        end
+
+        function get_model_o2()
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x[1:end](0) == [1, 0, 0]
+                ẋ(t) == [x₂(t), x₃(t), u(t)]
+                ∫(u(t)^2) → min
+            end
+        end
+
+        function get_model_o3()
+            n = 3
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x[1:n](0) == [1, 0, 0]
+                ẋ(t) == [x₂(t), x₃(t), u(t)]
+                ∫(u(t)^2) → min
+            end
+        end
+
+        # ====================================================================
+        # UNIT TESTS - Slices in constraints
+        # ====================================================================
+
+        Test.@testset "build() - Model with x[1:3](0) == x0 constraint" begin
+            o1 = get_model_o1()
+            Test.@test o1 isa OCP.Model
+            Test.@test OCP.state_dimension(o1) == 3
+            Test.@test OCP.control_dimension(o1) == 1
+        end
+
+        Test.@testset "build() - Model with x[1:end](0) == x0 constraint (broken)" begin
+            Test.@test_broken get_model_o2() isa OCP.Model
+        end
+
+        Test.@testset "build() - Model with x[1:n](0) == x0 constraint (broken)" begin
+            Test.@test_broken get_model_o3() isa OCP.Model
+        end
+
+        # ====================================================================
+        # INTEGRATION TESTS - Slices in constraints (discretisation + solve)
+        # ====================================================================
+
+        Test.@testset "discretise_exa with x[1:3](0) == x0" begin
+            o1 = get_model_o1()
+            m1 = discretise_exa(o1)
+            Test.@test m1 isa ExaModels.ExaModel
+        end
+
+        Test.@testset "discretise_exa with x[1:end](0) == x0 (broken)" begin
+            Test.@test_broken discretise_exa(get_model_o2()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "discretise_exa with x[1:n](0) == x0 (broken)" begin
+            Test.@test_broken discretise_exa(get_model_o3()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "solve with x[1:3](0) == x0" begin
+            o1 = get_model_o1()
+            m1 = discretise_exa(o1)
+            sol1 = MadNLP.madnlp(m1; tol=1e-6, print_level=MadNLP.ERROR)
+            Test.@test sol1.status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        Test.@testset "solve with x[1:end](0) == x0 (broken)" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_o2()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        Test.@testset "solve with x[1:n](0) == x0 (broken)" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_o3()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        # ====================================================================
+        # MODEL BUILDERS - Slices in dynamics
+        # ====================================================================
+        # These models test slice syntax in dynamics: ∂(x[...])(t) == ...
+        # - d1: ∂(x[1:3])(t)   → literal integer range (works for parsing, broken for discretisation)
+        # - d2: ∂(x[1:end])(t) → end keyword (broken - not yet supported)
+        # - d3: ∂(x[1:n])(t)   → variable n (works for parsing, broken for discretisation)
+
+        function get_model_d1()
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x[1:3](0) == [1, 0, 0]
+                ∂(x[1:3])(t) == [x₂(t), x₃(t), u(t)]
+                ∫(u(t)^2) → min
+            end
+        end
+
+        function get_model_d2()
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x[1:3](0) == [1, 0, 0]
+                ∂(x[1:end])(t) == [x₂(t), x₃(t), u(t)]
+                ∫(u(t)^2) → min
+            end
+        end
+
+        function get_model_d3()
+            n = 3
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x[1:3](0) == [1, 0, 0]
+                ∂(x[1:n])(t) == [x₂(t), x₃(t), u(t)]
+                ∫(u(t)^2) → min
+            end
+        end
+
+        # ====================================================================
+        # UNIT TESTS - Slices in dynamics
+        # ====================================================================
+
+        Test.@testset "build() - Model with ∂(x[1:3])(t) dynamics" begin
+            o = get_model_d1()
+            Test.@test o isa OCP.Model
+            Test.@test OCP.state_dimension(o) == 3
+            Test.@test OCP.control_dimension(o) == 1
+        end
+
+        Test.@testset "build() - Model with ∂(x[1:end])(t) dynamics (broken)" begin
+            Test.@test_broken get_model_d2() isa OCP.Model
+        end
+
+        Test.@testset "build() - Model with ∂(x[1:n])(t) dynamics (broken)" begin
+            o = get_model_d3()
+            Test.@test o isa OCP.Model
+            Test.@test OCP.state_dimension(o) == 3
+            Test.@test OCP.control_dimension(o) == 1
+        end
+
+        # ====================================================================
+        # INTEGRATION TESTS - Slices in dynamics (discretisation + solve)
+        # ====================================================================
+
+        Test.@testset "discretise_exa with ∂(x[1:3])(t) dynamics" begin
+            Test.@test_broken discretise_exa(get_model_d1()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "discretise_exa with ∂(x[1:end])(t) dynamics (broken)" begin
+            Test.@test_broken discretise_exa(get_model_d2()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "discretise_exa with ∂(x[1:n])(t) dynamics (broken)" begin
+            Test.@test_broken discretise_exa(get_model_d3()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "solve with ∂(x[1:3])(t) dynamics" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_d1()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        Test.@testset "solve with ∂(x[1:end])(t) dynamics (broken)" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_d2()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        Test.@testset "solve with ∂(x[1:n])(t) dynamics (broken)" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_d3()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        # ====================================================================
+        # MODEL BUILDERS - Slices in Mayer cost
+        # ====================================================================
+        # These models test slice syntax in Mayer cost: x[...](tf)
+        # - m1: x[1:3](tf)   → literal integer range (works for parsing, broken for discretisation)
+        # - m2: x[1:end](tf) → end keyword (works for parsing, broken for discretisation)
+        # - m3: x[1:n](tf)   → variable n (works for parsing, broken for discretisation)
+        function get_model_m1()
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x(0) == [1, 0, 0]
+                ∂(x)(t) == [x₂(t), x₃(t), u(t)]
+                sum(x[1:3](tf).^2) + ∫(u(t)^2) → min
+            end
+        end
+
+        function get_model_m2()
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x(0) == [1, 0, 0]
+                ∂(x)(t) == [x₂(t), x₃(t), u(t)]
+                sum(x[1:end](tf).^2) + ∫(u(t)^2) → min
+            end
+        end
+
+        function get_model_m3()
+            n = 3
+            return CTParser.@def begin
+                t ∈ [0, 1], time
+                x ∈ R^3, state
+                u ∈ R, control
+                x(0) == [1, 0, 0]
+                ∂(x)(t) == [x₂(t), x₃(t), u(t)]
+                sum(x[1:n](tf).^2) + ∫(u(t)^2) → min
+            end
+        end
+
+        # ====================================================================
+        # UNIT TESTS - Slices in Mayer cost
+        # ====================================================================
+
+        Test.@testset "build() - Model with x[1:3](tf) in Mayer cost" begin
+            o = get_model_m1()
+            Test.@test o isa OCP.Model
+            Test.@test OCP.state_dimension(o) == 3
+            Test.@test OCP.control_dimension(o) == 1
+        end
+
+        Test.@testset "build() - Model with x[1:end](tf) in Mayer cost" begin
+            o = get_model_m2()
+            Test.@test o isa OCP.Model
+            Test.@test OCP.state_dimension(o) == 3
+            Test.@test OCP.control_dimension(o) == 1
+        end
+
+        Test.@testset "build() - Model with x[1:n](tf) in Mayer cost" begin
+            o = get_model_m3()
+            Test.@test o isa OCP.Model
+            Test.@test OCP.state_dimension(o) == 3
+            Test.@test OCP.control_dimension(o) == 1
+        end
+
+        # ====================================================================
+        # INTEGRATION TESTS - Slices in Mayer cost (discretisation + solve)
+        # ====================================================================
+
+        Test.@testset "discretise_exa with x[1:3](tf) in Mayer cost" begin
+            Test.@test_broken discretise_exa(get_model_m1()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "discretise_exa with x[1:end](tf) in Mayer cost" begin
+            Test.@test_broken discretise_exa(get_model_m2()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "discretise_exa with x[1:n](tf) in Mayer cost" begin
+            Test.@test_broken discretise_exa(get_model_m3()) isa ExaModels.ExaModel
+        end
+
+        Test.@testset "solve with x[1:3](tf) in Mayer cost" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_m1()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        Test.@testset "solve with x[1:end](tf) in Mayer cost" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_m2()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+        Test.@testset "solve with x[1:n](tf) in Mayer cost" begin
+            Test.@test_broken MadNLP.madnlp(discretise_exa(get_model_m3()); tol=1e-6, print_level=MadNLP.ERROR).status == MadNLP.SOLVE_SUCCEEDED
+        end
+
+    end
+end
+
+end # module
+
+# CRITICAL: Redefine in outer scope for TestRunner
+test_slices() = TestSlices.test_slices()
\ No newline at end of file