diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 95c29c6..0f97b3c 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -8,6 +8,8 @@ on:
 permissions:
   checks: write
   contents: write
+  # deployments permission to deploy GitHub pages website
+  deployments: write
   pull-requests: write
 
 
@@ -85,3 +87,39 @@ jobs:
                 echo "Coverage Tests - ${{ steps.coverageComment.outputs.tests }}"
                 echo "Coverage Time - ${{ steps.coverageComment.outputs.time }}"
                 echo "Not Success Test Info - ${{ steps.coverageComment.outputs.notSuccessTestInfo }}"
+
+            - name: Run benchmarks
+              run: |
+                pytest -m benchmark --benchmark-json=./output.json
+
+            - name: Download previous benchmark data
+              uses: actions/cache@v4
+              with:
+                path: ./cache
+                key: ${{ runner.os }}-benchmark
+
+            - name: Publish benchmark results
+              uses: benchmark-action/github-action-benchmark@v1
+              if: github.event_name != 'pull_request'
+              with:
+                tool: 'pytest'
+                auto-push: true
+                comment-always: true
+                output-file-path: output.json
+                github-token: ${{ secrets.GITHUB_TOKEN }}
+                comment-on-alert: true
+                save-data-file: true
+                summary-always: true
+
+            - name: Comment on benchmark results without publishing
+              uses: benchmark-action/github-action-benchmark@v1
+              with:
+                tool: 'pytest'
+                auto-push: false
+                github-token: ${{ secrets.GITHUB_TOKEN }}
+                comment-always: true
+                output-file-path: output.json
+                comment-on-alert: false
+                save-data-file: true
+                summary-always: true
+                external-data-json-path: ./cache/benchmark-data.json
diff --git a/README.md b/README.md
index a571813..9fc6ede 100644
--- a/README.md
+++ b/README.md
@@ -401,6 +401,7 @@ Since the `EDTFField` and the `_earliest` and `_latest` field values are set aut
 
 ### Running tests
 - From `python-edtf`, run the unit tests: `pytest`
+- From `python-edtf`, run `pytest -m benchmark` to run the benchmarks (published [here]( https://ixc.github.io/python-edtf/dev/bench/))
 - From `python-edtf/edtf_django_tests`, run the integration tests: `python manage.py test edtf_integration`
 - To run CI locally, use `act`, e.g. `act pull_request` or `act --pull=false --container-architecture linux/amd64`. Some steps may require a Github PAT: `act pull_request --container-architecture linux/amd64 --pull=false -s GITHUB_TOKEN=<your PAT>`
 
diff --git a/dev-requirements.txt b/dev-requirements.txt
index 1e37df5..19242af 100644
--- a/dev-requirements.txt
+++ b/dev-requirements.txt
@@ -1,5 +1,7 @@
 -r requirements.txt  # Include all main requirements
 django>=4.2,<5.0
 pytest
+pytest-benchmark
+pytest-django
 ruff
 pre-commit
diff --git a/edtf/natlang/tests.py b/edtf/natlang/tests.py
index 78ecbc9..d2c43a5 100644
--- a/edtf/natlang/tests.py
+++ b/edtf/natlang/tests.py
@@ -185,3 +185,27 @@ def test_natlang(input_text, expected_output):
     assert (
         result == expected_output
     ), f"Failed for input: {input_text} - expected {expected_output}, got {result}"
+
+
+@pytest.mark.benchmark
+@pytest.mark.parametrize(
+    "input_text,expected_output",
+    [
+        ("23rd Dynasty", None),
+        ("January 2008", "2008-01"),
+        ("ca1860", "1860~"),
+        ("uncertain: approx 1862", "1862%"),
+        ("January", "XXXX-01"),
+        ("Winter 1872", "1872-24"),
+        ("before approx January 18 1928", "/1928-01-18~"),
+        ("birthday in 1872", "1872"),
+        ("1270 CE", "1270"),
+        ("2nd century bce", "-01XX"),
+        ("1858/1860", "[1858, 1860]"),
+    ],
+)
+def test_benchmark_natlang(benchmark, input_text, expected_output):
+    """
+    Benchmark selected natural language conversions
+    """
+    benchmark(text_to_edtf, input_text)
diff --git a/edtf/parser/grammar.py b/edtf/parser/grammar.py
index e6232c4..dc0f66d 100644
--- a/edtf/parser/grammar.py
+++ b/edtf/parser/grammar.py
@@ -1,3 +1,12 @@
+# ruff: noqa: E402 I001
+
+# It's recommended to `enablePackrat()` immediately after importing pyparsing
+# https://github.com/pyparsing/pyparsing/wiki/Performance-Tips
+
+import pyparsing
+
+pyparsing.ParserElement.enablePackrat()
+
 from pyparsing import (
     Combine,
     NotAny,
@@ -13,6 +22,7 @@
 )
 from pyparsing import Literal as L
 
+
 from edtf.parser.edtf_exceptions import EDTFParseException
 
 # (* ************************** Level 0 *************************** *)
diff --git a/edtf/parser/tests.py b/edtf/parser/tests.py
index 1ec7452..4932e95 100644
--- a/edtf/parser/tests.py
+++ b/edtf/parser/tests.py
@@ -216,6 +216,20 @@
     ("2001-34", ("2001-04-01", "2001-06-30")),
 )
 
+BENCHMARK_EXAMPLES = (
+    "2001-02-03",
+    "2008-12",
+    "2008",
+    "-0999",
+    "2004-01-01T10:10:10+05:00",
+    "-2005/-1999-02",
+    "/2006",
+    "?2004-%06",
+    "[1667, 1760-12]",
+    "Y3388E2S3",
+    "2001-29",
+)
+
 BAD_EXAMPLES = (
     # parentheses are not used for group qualification in the 2018 spec
     None,
@@ -340,3 +354,10 @@ def test_comparisons():
     assert d4 == d5
     assert d1 < d5
     assert d1 > d6
+
+
+@pytest.mark.benchmark
+@pytest.mark.parametrize("test_input", BENCHMARK_EXAMPLES)
+def test_benchmark_parser(benchmark, test_input):
+    """Benchmark parsing of selected EDTF strings."""
+    benchmark(parse, test_input)
diff --git a/pyproject.toml b/pyproject.toml
index 64579ae..56978fb 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -39,6 +39,7 @@ test = [
     "django>=4.2,<5.0",
     "pytest",
     "pytest-django",
+    "pytest-benchmark",
     "ruff",
     "pre-commit",
     "coverage",
@@ -81,8 +82,11 @@ legacy_tox_ini = """
 python_files = ["tests.py", "test_*.py", "*_test.py", "*_tests.py"]
 python_classes = ["Test*", "*Tests"]
 python_functions = ["test_*"]
-addopts = "--ignore=edtf_django_tests/ --cov=edtf"
-plugins = ["pytest_cov"]
+markers = [
+    "benchmark: mark a test as a benchmark",
+]
+addopts = "--ignore=edtf_django_tests/ --cov=edtf -m 'not benchmark'"
+plugins = ["pytest_cov", "pytest_benchmark"]
 
 [tool.coverage.run]
 # we run the edtf_integration tests but only care about them testing fields.py in the main package