diff --git a/.github/workflows/run_benchmarks.yml b/.github/workflows/run_benchmarks.yml
new file mode 100644
index 000000000..01dcdcc9b
--- /dev/null
+++ b/.github/workflows/run_benchmarks.yml
@@ -0,0 +1,80 @@
+name: Run performance benchmarks
+on:
+  pull_request:
+    types:
+      - labeled
+      - synchronize
+  workflow_dispatch:
+
+permissions:
+  contents: write
+  deployments: write
+  pull-requests: write
+
+jobs:
+  benchmark:
+    if: contains(github.event.label.name, 'benchmark') || github.event_name == 'workflow_dispatch'
+    name: Run pytest-benchmark
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Setup pixi environment
+        uses: prefix-dev/setup-pixi@v0.8.1
+        with:
+          pixi-version: v0.29.0
+          cache: true
+          auth-host: prefix.dev
+          auth-token: ${{ secrets.PREFIX_DEV_TOKEN }}
+
+      - name: Run benchmarks using pixi
+        run: pixi run benchmarks --env test-cpu
+
+      - name: Store benchmark result
+        uses: benchmark-action/github-action-benchmark@v1
+        with:
+          name: Python Benchmark with pytest-benchmark
+          tool: 'pytest'
+          output-file-path: tests/output.json
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+          auto-push: true
+          alert-threshold: '200%'
+          comment-on-alert: true
+          fail-on-alert: true
+
+      - name: Convert JSON to Markdown Table
+        id: convert_json
+        run: |
+          python3 - <<EOF
+          import json
+
+          # Load the JSON data
+          with open('tests/output.json', 'r') as file:
+              data = json.load(file)
+
+          # Extract keys and rows
+          keys = data[0].keys() if len(data) > 0 else []
+          rows = [[str(item[key]) for key in keys] for item in data]
+
+          # Create the Markdown table
+          table_header = '| ' + ' | '.join(keys) + ' |'
+          table_divider = '| ' + ' | '.join(['---'] * len(keys)) + ' |'
+          table_rows = ['| ' + ' | '.join(row) + ' |' for row in rows]
+          markdown_table = '\n'.join([table_header, table_divider] + table_rows)
+
+          # Save Markdown table as an output variable
+          print(f"::set-output name=markdown_table::{markdown_table}")
+          EOF
+
+      - name: Update PR Description with Benchmark Results
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          # Get the current PR description
+          PR_BODY=$(gh api repos/${{ github.repository }}/pulls/${{ github.event.number }} --jq .body)
+
+          # Append the benchmark results to the description
+          UPDATED_BODY="$PR_BODY\n\n## Benchmark Results\n${{ steps.convert_json.outputs.markdown_table }}"
+
+          # Update the PR description
+          gh api -X PATCH repos/${{ github.repository }}/pulls/${{ github.event.number }} -f body="$UPDATED_BODY"
diff --git a/pyproject.toml b/pyproject.toml
index 032f970a8..92f606df8 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -64,6 +64,7 @@ style = [
 testing = [
     "idyntree >= 12.2.1",
     "pytest >=6.0",
+    "pytest-benchmark",
     "pytest-icdiff",
     "robot-descriptions",
 ]
@@ -118,7 +119,7 @@ multi_line_output = 3
 profile = "black"
 
 [tool.pytest.ini_options]
-addopts = "-rsxX -v --strict-markers"
+addopts = "-rsxX -v --strict-markers --benchmark-skip"
 minversion = "6.0"
 testpaths = [
     "tests",
@@ -199,6 +200,7 @@ typing_extensions = "*"
 examples = { cmd = "jupyter notebook ./examples" }
 pipcheck = "pip check"
 test = { cmd = "pytest", depends_on = ["pipcheck"] }
+benchmark = { cmd = "pytest --benchmark-only --benchmark-json=output.json"}
 
 [tool.pixi.feature.test.dependencies]
 black = "24.*"
@@ -207,6 +209,7 @@ isort = "*"
 pip = "*"
 pre-commit = "*"
 pytest = "*"
+pytest-benchmark = "*"
 pytest-icdiff = "*"
 robot_descriptions = "*"
 
diff --git a/tests/benchmarks.py b/tests/benchmarks.py
new file mode 100644
index 000000000..1b4176b6a
--- /dev/null
+++ b/tests/benchmarks.py
@@ -0,0 +1,56 @@
+import pytest
+
+import jaxsim.api as js
+
+
+# Define the test functions for pytest-benchmark
+@pytest.mark.benchmark
+def test_free_floating_jacobian(
+    jaxsim_model_ergocub_reduced: js.model.JaxSimModel, benchmark
+):
+    model = jaxsim_model_ergocub_reduced
+    data = js.data.random_model_data(model=model)
+
+    # Warm-up call to avoid including compilation time
+    js.model.generalized_free_floating_jacobian(model=model, data=data)
+
+    benchmark(js.model.generalized_free_floating_jacobian, model, data)
+
+
+@pytest.mark.benchmark
+def test_free_floating_bias_forces(
+    jaxsim_model_ergocub_reduced: js.model.JaxSimModel, benchmark
+):
+    model = jaxsim_model_ergocub_reduced
+    data = js.data.random_model_data(model=model)
+
+    # Warm-up call to avoid including compilation time
+    js.model.free_floating_bias_forces(model=model, data=data)
+
+    benchmark(js.model.free_floating_bias_forces, model, data)
+
+
+@pytest.mark.benchmark
+def test_free_floating_mass_matrix(
+    jaxsim_model_ergocub_reduced: js.model.JaxSimModel, benchmark
+):
+    model = jaxsim_model_ergocub_reduced
+    data = js.data.random_model_data(model=model)
+
+    # Warm-up call to avoid including compilation time
+    js.model.free_floating_mass_matrix(model=model, data=data)
+
+    benchmark(js.model.free_floating_mass_matrix, model, data)
+
+
+@pytest.mark.benchmark
+def test_free_floating_jacobian_derivative(
+    jaxsim_model_ergocub_reduced: js.model.JaxSimModel, benchmark
+):
+    model = jaxsim_model_ergocub_reduced
+    data = js.data.random_model_data(model=model)
+
+    # Warm-up call to avoid including compilation time
+    js.model.generalized_free_floating_jacobian_derivative(model=model, data=data)
+
+    benchmark(js.model.generalized_free_floating_jacobian_derivative, model, data)