diff --git a/python/src/main.cpp b/python/src/main.cpp
index d94b2ecc..94e4b134 100644
--- a/python/src/main.cpp
+++ b/python/src/main.cpp
@@ -534,7 +534,7 @@ PYBIND11_MODULE(kp, m)
             if (spec_consts.dtype().is(py::dtype::of<std::float_t>())) {
                 std::vector<float> specConstsVec(
                   (float*)specInfo.ptr, ((float*)specInfo.ptr) + specInfo.size);
-                if (spec_consts.dtype().is(py::dtype::of<std::float_t>())) {
+                if (push_consts.dtype().is(py::dtype::of<std::float_t>())) {
                     std::vector<float> pushConstsVec((float*)pushInfo.ptr,
                                                      ((float*)pushInfo.ptr) +
                                                        pushInfo.size);
@@ -543,7 +543,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specConstsVec,
                                           pushConstsVec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::int32_t>())) {
                     std::vector<int32_t> pushConstsVec(
                       (int32_t*)pushInfo.ptr,
@@ -553,7 +553,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specConstsVec,
                                           pushConstsVec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::uint32_t>())) {
                     std::vector<uint32_t> pushConstsVec(
                       (uint32_t*)pushInfo.ptr,
@@ -563,7 +563,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specConstsVec,
                                           pushConstsVec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::double_t>())) {
                     std::vector<double> pushConstsVec((double*)pushInfo.ptr,
                                                       ((double*)pushInfo.ptr) +
@@ -578,7 +578,7 @@ PYBIND11_MODULE(kp, m)
                 std::vector<int32_t> specconstsvec((int32_t*)specInfo.ptr,
                                                    ((int32_t*)specInfo.ptr) +
                                                      specInfo.size);
-                if (spec_consts.dtype().is(py::dtype::of<std::float_t>())) {
+                if (push_consts.dtype().is(py::dtype::of<std::float_t>())) {
                     std::vector<float> pushconstsvec((float*)pushInfo.ptr,
                                                      ((float*)pushInfo.ptr) +
                                                        pushInfo.size);
@@ -587,7 +587,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::int32_t>())) {
                     std::vector<int32_t> pushconstsvec(
                       (int32_t*)pushInfo.ptr,
@@ -597,7 +597,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::uint32_t>())) {
                     std::vector<uint32_t> pushconstsvec(
                       (uint32_t*)pushInfo.ptr,
@@ -607,7 +607,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::double_t>())) {
                     std::vector<double> pushconstsvec((double*)pushInfo.ptr,
                                                       ((double*)pushInfo.ptr) +
@@ -622,7 +622,7 @@ PYBIND11_MODULE(kp, m)
                 std::vector<uint32_t> specconstsvec((uint32_t*)specInfo.ptr,
                                                     ((uint32_t*)specInfo.ptr) +
                                                       specInfo.size);
-                if (spec_consts.dtype().is(py::dtype::of<std::float_t>())) {
+                if (push_consts.dtype().is(py::dtype::of<std::float_t>())) {
                     std::vector<float> pushconstsvec((float*)pushInfo.ptr,
                                                      ((float*)pushInfo.ptr) +
                                                        pushInfo.size);
@@ -631,7 +631,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::int32_t>())) {
                     std::vector<int32_t> pushconstsvec(
                       (int32_t*)pushInfo.ptr,
@@ -641,7 +641,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::uint32_t>())) {
                     std::vector<uint32_t> pushconstsvec(
                       (uint32_t*)pushInfo.ptr,
@@ -651,7 +651,7 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
+                } else if (push_consts.dtype().is(
                              py::dtype::of<std::double_t>())) {
                     std::vector<double> pushconstsvec((double*)pushInfo.ptr,
                                                       ((double*)pushInfo.ptr) +
@@ -666,7 +666,7 @@ PYBIND11_MODULE(kp, m)
                 std::vector<double> specconstsvec((double*)specInfo.ptr,
                                                   ((double*)specInfo.ptr) +
                                                     specInfo.size);
-                if (spec_consts.dtype().is(py::dtype::of<std::float_t>())) {
+                if (push_consts.dtype().is(py::dtype::of<std::float_t>())) {
                     std::vector<float> pushconstsvec((float*)pushInfo.ptr,
                                                      ((float*)pushInfo.ptr) +
                                                        pushInfo.size);
@@ -675,31 +675,25 @@ PYBIND11_MODULE(kp, m)
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
-                             py::dtype::of<std::int32_t>())) {
-                    std::vector<float> pushconstsvec((int32_t*)pushInfo.ptr,
-                                                     ((int32_t*)pushInfo.ptr) +
-                                                       pushInfo.size);
+                } else if (push_consts.dtype().is(py::dtype::of<std::int32_t>())) {
+                    std::vector<int32_t> pushconstsvec((int32_t*)pushInfo.ptr,
+                                           ((int32_t*)pushInfo.ptr) + pushInfo.size);
                     return self.algorithm(tensors,
                                           spirvVec,
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
-                             py::dtype::of<std::uint32_t>())) {
-                    std::vector<float> pushconstsvec((uint32_t*)pushInfo.ptr,
-                                                     ((uint32_t*)pushInfo.ptr) +
-                                                       pushInfo.size);
+                } else if (push_consts.dtype().is(py::dtype::of<std::uint32_t>())) {
+                    std::vector<uint32_t> pushconstsvec((uint32_t*)pushInfo.ptr,
+                                            ((uint32_t*)pushInfo.ptr) + pushInfo.size);
                     return self.algorithm(tensors,
                                           spirvVec,
                                           workgroup,
                                           specconstsvec,
                                           pushconstsvec);
-                } else if (spec_consts.dtype().is(
-                             py::dtype::of<std::double_t>())) {
-                    std::vector<float> pushconstsvec((double*)pushInfo.ptr,
-                                                     ((double*)pushInfo.ptr) +
-                                                       pushInfo.size);
+                } else if (push_consts.dtype().is(py::dtype::of<std::double_t>())) {
+                    std::vector<double> pushconstsvec((double*)pushInfo.ptr,
+                                          ((double*)pushInfo.ptr) + pushInfo.size);
                     return self.algorithm(tensors,
                                           spirvVec,
                                           workgroup,
diff --git a/python/test/test_consts.py b/python/test/test_consts.py
new file mode 100644
index 00000000..44783289
--- /dev/null
+++ b/python/test/test_consts.py
@@ -0,0 +1,259 @@
+import kp
+import numpy as np
+import pytest
+
+from .utils import compile_source
+
+workgroup = (3, 1, 1)
+
+def test_pushconsts_int32_spec_const_int32():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const int spec_val = 42;
+        layout(push_constant) uniform PushConsts {
+            int value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([1, 2, 3], dtype=np.int32)
+    spec_const = np.array([5], dtype=np.int32)  # Will override spec_val to 5
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_int32_spec_const_uint32():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const uint spec_val = 0u;
+        layout(push_constant) uniform PushConsts {
+            int value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([4, 5, 6], dtype=np.int32)
+    spec_const = np.array([7], dtype=np.uint32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_int32_spec_const_float():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const float spec_val = 0.0;
+        layout(push_constant) uniform PushConsts {
+            int value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([7, 8, 9], dtype=np.int32)
+    spec_const = np.array([3.3], dtype=np.float32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_uint32_spec_const_int32():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const int spec_val = 0;
+        layout(push_constant) uniform PushConsts {
+            uint value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([1, 2, 3], dtype=np.uint32)
+    spec_const = np.array([4], dtype=np.int32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_uint32_spec_const_uint32():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const uint spec_val = 0u;
+        layout(push_constant) uniform PushConsts {
+            uint value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([4, 5, 6], dtype=np.uint32)
+    spec_const = np.array([8], dtype=np.uint32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_uint32_spec_const_float():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const float spec_val = 0.0;
+        layout(push_constant) uniform PushConsts {
+            uint value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([7, 8, 9], dtype=np.uint32)
+    spec_const = np.array([3.3], dtype=np.float32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_float_spec_const_int32():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const int spec_val = 42;
+        layout(push_constant) uniform PushConsts {
+            float value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([1.1, 2.2, 3.3], dtype=np.float32)
+    spec_const = np.array([11], dtype=np.int32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_float_spec_const_uint32():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const uint spec_val = 0u;
+        layout(push_constant) uniform PushConsts {
+            float value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([4.4, 5.5, 6.6], dtype=np.float32)
+    spec_const = np.array([2], dtype=np.uint32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])
+
+def test_pushconsts_float_spec_const_float():
+    shader = """
+        #version 450
+        layout(constant_id = 0) const float spec_val = 0.0;
+        layout(push_constant) uniform PushConsts {
+            float value[3];
+        } pc;
+        layout(set = 0, binding = 0) buffer Output {
+            float outData[];
+        };
+        void main() {
+            uint idx = gl_GlobalInvocationID.x;
+            outData[idx] = float(pc.value[idx]) + float(spec_val);
+        }
+    """
+    spirv = compile_source(shader)
+    mgr = kp.Manager()
+    arr_out = np.array([0.0, 0.0, 0.0], dtype=np.float32)
+    tensor_out = mgr.tensor_t(arr_out)
+    push_consts = np.array([7.7, 8.8, 9.9], dtype=np.float32)
+    spec_const = np.array([1.1], dtype=np.float32)
+    algo = mgr.algorithm([tensor_out], spirv, workgroup, spec_const, push_consts)
+    (mgr.sequence()
+        .record(kp.OpAlgoDispatch(algo))
+        .record(kp.OpSyncLocal([tensor_out]))
+        .eval())
+    assert np.array_equal(tensor_out.data(), push_consts.astype(np.float32) + spec_const.astype(np.float32)[0])