diff --git a/.editorconfig b/.editorconfig
deleted file mode 100644
index d4a2c44..0000000
--- a/.editorconfig
+++ /dev/null
@@ -1,21 +0,0 @@
-# http://editorconfig.org
-
-root = true
-
-[*]
-indent_style = space
-indent_size = 4
-trim_trailing_whitespace = true
-insert_final_newline = true
-charset = utf-8
-end_of_line = lf
-
-[*.bat]
-indent_style = tab
-end_of_line = crlf
-
-[LICENSE]
-insert_final_newline = false
-
-[Makefile]
-indent_style = tab
diff --git a/.github/ISSUE_TEMPLATE.md b/.github/ISSUE_TEMPLATE.md
deleted file mode 100644
index 1a15181..0000000
--- a/.github/ISSUE_TEMPLATE.md
+++ /dev/null
@@ -1,15 +0,0 @@
-* mt2 version:
-* Python version:
-* Operating System:
-
-### Description
-
-Describe what you were trying to get done.
-Tell us what happened, what went wrong, and what you expected to happen.
-
-### What I Did
-
-```
-Paste the command(s) you ran and the output.
-If there was a crash, please include the traceback here.
-```
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 4b2807c..794101c 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -1,11 +1,17 @@
 name: Build
 
-on: [push, pull_request]
+on:
+    pull_request:
+    push:
+        branches: [main]
+        tags: ["*"]
+
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}
   cancel-in-progress: true
 
+
 jobs:
     test:
         runs-on: ubuntu-latest
@@ -15,22 +21,26 @@ jobs:
 
         steps:
             - uses: actions/checkout@v4
-            - name: Set up Python ${{ matrix.python-version }}
-              uses: actions/setup-python@v5
-              with:
-                  python-version: ${{ matrix.python-version }}
-            - name: Install dependencies
+
+            - name: Install uv
+              run: curl -LsSf https://astral.sh/uv/install.sh | sh
+
+            - name: Install & activate
               run: |
-                  python -m pip install --upgrade pip
-                  pip install -U -r requirements_dev.txt
-                  pip install tox tox-gh-actions
-            - name: Test with tox numpy>=2
-              run: tox -e latestnpy
-            - name: Test with tox numpy<2
-              run: tox -e oldestnpy
+                uv sync --python ${{ matrix.python-version }}
+                . .venv/bin/activate
+                echo PATH=$PATH >> $GITHUB_ENV
+
+            - name: Test
+              run: python -m unittest discover tests
+
+            - name: Test with oldest-supported-numpy
+              run: |
+                uv pip install oldest-supported-numpy 
+                python -m unittest discover tests
 
     build_wheels:
-        name: Build wheel for ${{ matrix.os }}-cp${{ matrix.python }}-${{ matrix.arch }}
+        name: Build wheel for ${{ matrix.os }}-${{ matrix.build }}${{ matrix.python }}-${{ matrix.arch }}
         runs-on: ${{ matrix.os }}
         strategy:
             # Ensure that a wheel builder finishes even if another fails
@@ -51,18 +61,20 @@ jobs:
             - name: Checkout mt2
               uses: actions/checkout@v4
 
-            # Used to host cibuildwheel
-            - name: Setup Python
-              uses: actions/setup-python@v5
+            # Needed within cibuildwheel
+            - name: Install uv
+              run: curl -LsSf https://astral.sh/uv/install.sh | sh
 
             - uses: pypa/cibuildwheel@v2.20.0
               env:
+                CIBW_BUILD_FRONTEND: "build[uv]"
                 CIBW_BUILD: "${{ matrix.build }}${{ matrix.python }}*"
                 CIBW_ARCHS: ${{ matrix.arch }}
+                CIBW_TEST_COMMAND: python -m unittest discover -t {project} -s {project}/tests
 
             - uses: actions/upload-artifact@v4
-              if: github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/v')
               with:
+                  name: "artifact-${{ matrix.os }}-${{ matrix.build }}-${{ matrix.python }}-${{ matrix.arch }}"
                   path: ./wheelhouse/*.whl
 
     build_sdist:
@@ -71,20 +83,20 @@ jobs:
         steps:
             - uses: actions/checkout@v4
 
-            - uses: actions/setup-python@v5
-              name: Install Python
-              with:
-                  python-version: "3.12"
+            - name: Install uv
+              run: curl -LsSf https://astral.sh/uv/install.sh | sh
 
-            - name: Install numpy and setuptools
-              run: python -m pip install numpy setuptools
+            - name: Install environment & setuptools
+              run: |
+                uv sync
+                uv pip install setuptools
 
             - name: Build sdist
-              run: python setup.py sdist
+              run: uv run python setup.py sdist
 
             - uses: actions/upload-artifact@v4
-              if: github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/v')
               with:
+                  name: artifact-source
                   path: dist/*.tar.gz
 
     pass:
@@ -94,7 +106,7 @@ jobs:
         - run: echo "All jobs passed"
 
     upload_pypi:
-        needs: [build_wheels, build_sdist, test]
+        needs: [pass]
         runs-on: ubuntu-latest
         # upload to PyPI on every tag starting with 'v'
         if: github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/v')
@@ -103,8 +115,8 @@ jobs:
         steps:
             - uses: actions/download-artifact@v4
               with:
-                  name: artifact
                   path: dist
+                  merge-multiple: true
 
             - uses: pypa/gh-action-pypi-publish@release/v1
               with:
diff --git a/.gitignore b/.gitignore
index 89821a8..6aeca48 100644
--- a/.gitignore
+++ b/.gitignore
@@ -28,55 +28,8 @@ wheels/
 .installed.cfg
 *.egg
 
-# PyInstaller
-#  Usually these files are written by a python script from a template
-#  before PyInstaller builds the exe, so as to inject date/other infos into it.
-*.manifest
-*.spec
-
-# Installer logs
-pip-log.txt
-pip-delete-this-directory.txt
-
-# Unit test / coverage reports
-htmlcov/
-.tox/
-.coverage
-.coverage.*
-.cache
-nosetests.xml
-coverage.xml
-*.cover
-.hypothesis/
-.pytest_cache/
-
-# Translations
-*.mo
-*.pot
-
-# pyenv
-.python-version
-
-# celery beat schedule file
-celerybeat-schedule
-
-# SageMath parsed files
-*.sage.py
-
-# dotenv
-.env
-
 # virtualenv
 .venv
-venv/
-ENV/
-
-# Rope project settings
-.ropeproject
-
-# mypy
-.mypy_cache/
 
-# IDE settings
-.idea/
-.vscode/
+# package manager lock file
+uv.lock
diff --git a/AUTHORS.rst b/AUTHORS.rst
deleted file mode 100644
index dc38a60..0000000
--- a/AUTHORS.rst
+++ /dev/null
@@ -1,7 +0,0 @@
-=======
-Credits
-=======
-
-* Christopher Lester <lester@hep.phy.cam.ac.uk>: Original C++ implementation of mT2.
-* Rupert Tombs: Current C++ implementation used in this package.
-* Thomas Gillam <tpgillam@googlemail.com>: Python packaging
diff --git a/CONTRIBUTING.rst b/CONTRIBUTING.rst
deleted file mode 100644
index 752838a..0000000
--- a/CONTRIBUTING.rst
+++ /dev/null
@@ -1,127 +0,0 @@
-.. highlight:: shell
-
-============
-Contributing
-============
-
-Contributions are welcome, and they are greatly appreciated! Every little bit
-helps, and credit will always be given.
-
-You can contribute in many ways:
-
-Types of Contributions
-----------------------
-
-Report Bugs
-~~~~~~~~~~~
-
-Report bugs at https://github.com/tpgillam/mt2/issues.
-
-If you are reporting a bug, please include:
-
-* Your operating system name and version.
-* Any details about your local setup that might be helpful in troubleshooting.
-* Detailed steps to reproduce the bug.
-
-Fix Bugs
-~~~~~~~~
-
-Look through the GitHub issues for bugs. Anything tagged with "bug" and "help
-wanted" is open to whoever wants to implement it.
-
-Implement Features
-~~~~~~~~~~~~~~~~~~
-
-Look through the GitHub issues for features. Anything tagged with "enhancement"
-and "help wanted" is open to whoever wants to implement it.
-
-Write Documentation
-~~~~~~~~~~~~~~~~~~~
-
-mt2 could always use more documentation, whether as part of the
-official mt2 docs, in docstrings, or even on the web in blog posts,
-articles, and such.
-
-Submit Feedback
-~~~~~~~~~~~~~~~
-
-The best way to send feedback is to file an issue at https://github.com/tpgillam/mt2/issues.
-
-If you are proposing a feature:
-
-* Explain in detail how it would work.
-* Keep the scope as narrow as possible, to make it easier to implement.
-* Remember that this is a volunteer-driven project, and that contributions
-  are welcome :)
-
-Get Started!
-------------
-
-Ready to contribute? Here's how to set up `mt2` for local development.
-
-1. Fork the `mt2` repo on GitHub.
-2. Clone your fork locally::
-
-    $ git clone git@github.com:your_name_here/mt2.git
-
-3. Install your local copy into a virtualenv. Assuming you have virtualenvwrapper installed, this is how you set up your fork for local development::
-
-    $ mkvirtualenv mt2
-    $ cd mt2/
-    $ python setup.py develop
-
-4. Create a branch for local development::
-
-    $ git checkout -b name-of-your-bugfix-or-feature
-
-   Now you can make your changes locally.
-
-5. When you're done making changes, check that your changes pass flake8 and the
-   tests, including testing other Python versions with tox::
-
-    $ flake8 mt2 tests
-    $ python setup.py test or pytest
-    $ tox
-
-   To get flake8 and tox, just pip install them into your virtualenv.
-
-6. Commit your changes and push your branch to GitHub::
-
-    $ git add .
-    $ git commit -m "Your detailed description of your changes."
-    $ git push origin name-of-your-bugfix-or-feature
-
-7. Submit a pull request through the GitHub website.
-
-Pull Request Guidelines
------------------------
-
-Before you submit a pull request, check that it meets these guidelines:
-
-1. The pull request should include tests.
-2. If the pull request adds functionality, the docs should be updated. Put
-   your new functionality into a function with a docstring, and add the
-   feature to the list in README.rst.
-3. The pull request should work across all supported platforms and Python versions.
-   Check that the continuous integration tests pass.
-
-Tips
-----
-
-To run a subset of tests::
-
-$ pytest tests.test_mt2
-
-
-Deploying
----------
-
-A reminder for the maintainers on how to deploy.
-Make sure all your changes are committed (including an entry in HISTORY.rst).
-Then run::
-
-$ bump2version patch # possible: major / minor / patch
-$ git push
-$ git push --tags
-
-Travis will then deploy to PyPI if tests pass.
diff --git a/HISTORY.rst b/HISTORY.rst
index b9a841c..ffaf498 100644
--- a/HISTORY.rst
+++ b/HISTORY.rst
@@ -6,6 +6,7 @@ History
 ------------------
 
 * Move support to Python 3.9-3.12. Support numpy 2. Thanks to @lgray
+* Various build system & package modernisation.
 
 1.2.0 (2021-05-05)
 ------------------
diff --git a/MANIFEST.in b/MANIFEST.in
deleted file mode 100644
index 2e98c08..0000000
--- a/MANIFEST.in
+++ /dev/null
@@ -1,11 +0,0 @@
-include AUTHORS.rst
-include CONTRIBUTING.rst
-include HISTORY.rst
-include LICENSE
-include README.rst
-include pyproject.toml
-
-recursive-include src *
-recursive-include tests *
-recursive-exclude * __pycache__
-recursive-exclude * *.py[co]
diff --git a/Makefile b/Makefile
index 7fffe8a..0399c58 100644
--- a/Makefile
+++ b/Makefile
@@ -1,74 +1,29 @@
-.PHONY: clean clean-test clean-pyc clean-build help
-.DEFAULT_GOAL := help
-
-define BROWSER_PYSCRIPT
-import os, webbrowser, sys
-
-from urllib.request import pathname2url
-
-webbrowser.open("file://" + pathname2url(os.path.abspath(sys.argv[1])))
-endef
-export BROWSER_PYSCRIPT
-
-define PRINT_HELP_PYSCRIPT
-import re, sys
-
-for line in sys.stdin:
-	match = re.match(r'^([a-zA-Z_-]+):.*?## (.*)$$', line)
-	if match:
-		target, help = match.groups()
-		print("%-20s %s" % (target, help))
-endef
-export PRINT_HELP_PYSCRIPT
-
-BROWSER := python -c "$$BROWSER_PYSCRIPT"
-
-help:
-	@python -c "$$PRINT_HELP_PYSCRIPT" < $(MAKEFILE_LIST)
-
-clean: clean-build clean-pyc clean-test ## remove all build, test, coverage and Python artifacts
-
-clean-build: ## remove build artifacts
-	rm -fr build/
-	rm -fr dist/
-	rm -fr .eggs/
-	find . -name '*.egg-info' -exec rm -fr {} +
-	find . -name '*.egg' -exec rm -f {} +
-
-clean-pyc: ## remove Python file artifacts
-	find . -name '*.pyc' -exec rm -f {} +
-	find . -name '*.pyo' -exec rm -f {} +
-	find . -name '*~' -exec rm -f {} +
+.PHONY: install
+install:
+	uv sync
+
+.PHONY: clean
+clean: clean-build clean-pyc clean-venv
+
+.PHONY: clean-build
+clean-build:
+	rm -rf build/
+	rm -rf dist/
+	rm -rf .eggs/
+	find . -name '*.egg-info' -delete
+	find . -name '*.egg' -delete
+
+.PHONY: clean-pyc
+clean-pyc:
+	find . -name '*.pyc' -delete
+	find . -name '*.pyo' -delete
 	find . -name '__pycache__' -exec rm -fr {} +
 
-clean-test: ## remove test and coverage artifacts
-	rm -fr .tox/
-	rm -f .coverage
-	rm -fr htmlcov/
-	rm -fr .pytest_cache
-
-lint: ## check style with flake8
-	flake8 mt2 tests
-
-test: ## run tests quickly with the default Python
-	pytest
-
-test-all: ## run tests on every Python version with tox
-	tox
-
-coverage: ## check code coverage quickly with the default Python
-	coverage run --source mt2 -m pytest
-	coverage report -m
-	coverage html
-	$(BROWSER) htmlcov/index.html
-
-release: dist ## package and upload a release
-	twine upload dist/*
-
-dist: clean ## builds source and wheel package
-	python setup.py sdist
-	python setup.py bdist_wheel
-	ls -l dist
+.PHONY: clean-venv
+clean-venv:
+	rm -f uv.lock
+	rm -rf .venv
 
-install: clean ## install the package to the active Python's site-packages
-	python setup.py install
+.PHONY: test
+test: install
+	uv run --locked python -m unittest discover tests
diff --git a/README.rst b/README.rst
index aa8c867..b5b7acf 100644
--- a/README.rst
+++ b/README.rst
@@ -5,11 +5,12 @@ mt2
 .. image:: https://img.shields.io/pypi/v/mt2.svg
         :target: https://pypi.python.org/pypi/mt2
 
+.. image:: https://img.shields.io/pypi/pyversions/mt2.svg
+        :target: https://pypi.python.org/pypi/mt2
+
 .. image:: https://github.com/tpgillam/mt2/workflows/Build/badge.svg?branch=master
         :target: https://github.com/tpgillam/mt2/actions?query=workflow%3ABuild
 
-.. image:: https://img.shields.io/badge/code%20style-black-000000.svg
-    :target: https://github.com/psf/black
 
 This package may be used to evaluate MT2 in all its variants.
 This includes both symmetric and asymmetric MT2.
@@ -140,3 +141,10 @@ A list of alternative implementations of the MT2 calculation can be found here:
 https://www.hep.phy.cam.ac.uk/~lester/mt2/#Alternatives
 
 In Python, the other wrapper of the same algorithm known to the authors is by Nikolai Hartmann, here: https://gitlab.cern.ch/nihartma/pymt2
+
+
+Authors
+-------
+* @kesterlester: Original C++ implementation of mT2.
+* @rupt: Current C++ implementation used in this package.
+* @tpgillam: Python packaging
diff --git a/mt2/__init__.py b/mt2/__init__.py
index 62e2a0f..026b10f 100644
--- a/mt2/__init__.py
+++ b/mt2/__init__.py
@@ -4,8 +4,6 @@
 
 from _mt2 import mt2_lester_ufunc, mt2_tombs_ufunc  # pyright: ignore [reportMissingImports]
 
-__author__ = "Thomas Gillam"
-__email__ = "tpgillam@googlemail.com"
 __version__ = "1.2.1"
 
 __all__ = ["mt2", "mt2_arxiv", "mt2_ufunc"]
diff --git a/pyproject.toml b/pyproject.toml
index 1d95a0b..db6faf9 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,33 @@
-[build-system]
-requires = [
-    "setuptools",
-    "wheel",
-    "numpy>=1.19.3",
+[project]
+name = "mt2"
+version = "1.2.1"
+description = "Stransverse mass computation as a numpy ufunc."
+authors = [
+    { name = "Tom Gillam", email = "tpgillam@googlemail.com" },
+    { name = "Rupert Tombs" },
+    { name = "Christopher Lester" },
+]
+readme = "README.rst"
+license = { file = "LICENSE" }
+classifiers = [
+    "Development Status :: 5 - Production/Stable",
+    "Intended Audience :: Developers",
+    "License :: OSI Approved :: MIT License",
+    "Natural Language :: English",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
 ]
+keywords = ["mt2"]
+dependencies = ["numpy>=1.19.3"]
+requires-python = ">= 3.9"
+urls = { Homepage = "https://github.com/tpgillam/mt2" }
+
+[build-system]
+requires = ["setuptools>=61.0", "wheel", "numpy"]
+build-backend = "setuptools.build_meta"
+
+[tool.setuptools.package-data]
+mt2 = ["src/*.h", "src/*.cpp"]
diff --git a/requirements_dev.txt b/requirements_dev.txt
deleted file mode 100644
index 2137fe4..0000000
--- a/requirements_dev.txt
+++ /dev/null
@@ -1,12 +0,0 @@
-pip
-numpy>=1.19.3
-bump2version
-wheel
-watchdog
-flake8
-tox
-coverage
-twine
-
-pytest
-pytest-runner
diff --git a/setup.cfg b/setup.cfg
deleted file mode 100644
index 31ad82b..0000000
--- a/setup.cfg
+++ /dev/null
@@ -1,2 +0,0 @@
-[aliases]
-test = pytest
diff --git a/setup.py b/setup.py
index 392ba58..fa1f817 100644
--- a/setup.py
+++ b/setup.py
@@ -4,66 +4,24 @@
 
 __version__ = "1.2.1"
 
-with open("README.rst") as readme_file:
-    readme = readme_file.read()
-
-with open("HISTORY.rst") as history_file:
-    history = history_file.read()
-
-ext_modules = [
-    Extension(
-        "_mt2",
-        ["src/main.cpp"],
-        define_macros=[
-            # Pass in the version info so we can expose it in the extension.
-            ("VERSION_INFO", __version__),
-            # For reasons explained in lester_mt2_bisect_v7.h, we need to manually
-            # enable some inlining optimisations.
-            ("ENABLE_INLINING", "1"),
-            # Copyright printing is disabled here, since we include the necessary
-            # citation information elsewhere.
-            ("DISABLE_COPYRIGHT_PRINTING", "1"),
-        ],
-        include_dirs=[numpy.get_include()],
-        language="c++",
-        extra_compile_args=["-std=c++11"],
-    ),
-]
-
-setup_requirements = [
-    "pytest-runner",
-]
-test_requirements = [
-    "pytest>=3",
-]
-
 setup(
-    author="Thomas Gillam",
-    author_email="tpgillam@googlemail.com",
-    classifiers=[
-        "Development Status :: 5 - Production/Stable",
-        "Intended Audience :: Developers",
-        "License :: OSI Approved :: MIT License",
-        "Natural Language :: English",
-        "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.9",
-        "Programming Language :: Python :: 3.10",
-        "Programming Language :: Python :: 3.11",
-        "Programming Language :: Python :: 3.12",
+    ext_modules=[
+        Extension(
+            "_mt2",
+            ["src/main.cpp"],
+            define_macros=[
+                # Pass in the version info so we can expose it in the extension.
+                ("VERSION_INFO", __version__),
+                # For reasons explained in lester_mt2_bisect_v7.h, we need to manually
+                # enable some inlining optimisations.
+                ("ENABLE_INLINING", "1"),
+                # Copyright printing is disabled here, since we include the necessary
+                # citation information elsewhere.
+                ("DISABLE_COPYRIGHT_PRINTING", "1"),
+            ],
+            include_dirs=[numpy.get_include()],
+            language="c++",
+            extra_compile_args=["-std=c++11"],
+        ),
     ],
-    description="Stransverse mass computation as a numpy ufunc.",
-    install_requires=["numpy>=1.19.3"],
-    license="MIT license",
-    long_description=readme + "\n\n" + history,
-    include_package_data=True,
-    keywords="mt2",
-    name="mt2",
-    packages=["mt2"],
-    setup_requires=setup_requirements,
-    test_suite="tests",
-    tests_require=test_requirements,
-    url="https://github.com/tpgillam/mt2",
-    version=__version__,
-    ext_modules=ext_modules,
-    zip_safe=False,
 )
diff --git a/src/main.cpp b/src/main.cpp
index 979e2df..6d649cc 100644
--- a/src/main.cpp
+++ b/src/main.cpp
@@ -13,6 +13,7 @@
 #define STRINGIFY(x) #x
 #define MACRO_STRINGIFY(x) STRINGIFY(x)
 
+
 static void mt2_lester_ufunc(
     char **args,
 // const-correctness was introduced in numpy 1.19, but retain backward compatibility.
diff --git a/tests/test_mt2.py b/tests/test_mt2.py
index c059855..ff5b5f5 100644
--- a/tests/test_mt2.py
+++ b/tests/test_mt2.py
@@ -1,98 +1,114 @@
 """Tests for `mt2` package."""
+
 import math
 import random
+import unittest
 
-import pytest
+import numpy
 
 from mt2 import mt2, mt2_arxiv
 
 
-def test_simple_example():
-    for mt2_impl in (mt2, mt2_arxiv):
-        computed_val = mt2_impl(100, 410, 20, 150, -210, -300, -200, 280, 100, 100)
-        assert computed_val == pytest.approx(412.628)
-
-
-def test_near_massless():
-    # This test is based on Fig 5 of https://arxiv.org/pdf/1411.4312.pdf
-    m_vis_a = 0
-    px_a = -42.017340486
-    py_a = -146.365340528
-
-    m_vis_b = 0.087252259
-    px_b = -9.625614206
-    py_b = 145.757295514
-
-    px_miss = -16.692279406
-    py_miss = -14.730240471
-
-    chi_a = 0
-    chi_b = 0
+class TestMt2(unittest.TestCase):
+    def test_simple_example(self):
+        for mt2_impl in (mt2, mt2_arxiv):
+            computed_val = mt2_impl(100, 410, 20, 150, -210, -300, -200, 280, 100, 100)
+            self.assertAlmostEqual(computed_val, 412.627668458219)
 
-    for mt2_impl in (mt2, mt2_arxiv):
-        computed_val = mt2_impl(
-            m_vis_a, px_a, py_a, m_vis_b, px_b, py_b, px_miss, py_miss, chi_a, chi_b
-        )
-        assert computed_val == pytest.approx(0.09719971)
+    def test_near_massless(self):
+        # This test is based on Fig 5 of https://arxiv.org/pdf/1411.4312.pdf
+        m_vis_a = 0
+        px_a = -42.017340486
+        py_a = -146.365340528
 
+        m_vis_b = 0.087252259
+        px_b = -9.625614206
+        py_b = 145.757295514
 
-def test_collinear_endpoint_cases():
-    random.seed(0)  # If the test fails, we want to be able to repeat the test!
-    for i in range(10000):
-        m_vis_a = random.uniform(0, 10)
-        m_vis_b = random.uniform(0, 10)
-        m_invis_a = random.uniform(0, 10)
-        m_invis_b = random.uniform(0, 10)
+        px_miss = -16.692279406
+        py_miss = -14.730240471
 
-        # Addition of random positive number prevents possibility of division
-        # by zero at the locations below marked "MOO"
-        m_parent = max(m_vis_a + m_invis_a, m_vis_b + m_invis_b) + random.uniform(
-            0.1, 10
-        )
-        p_parent_a = random.uniform(0, 10)
-        p_parent_b = random.uniform(0, 10)
-        e_parent_a = math.sqrt(p_parent_a ** 2 + m_parent ** 2)
-        e_parent_b = math.sqrt(p_parent_b ** 2 + m_parent ** 2)
-        beta_a = p_parent_a / e_parent_a  # MOO
-        beta_b = p_parent_b / e_parent_b  # MOO
-        gamma_a = 1.0 / math.sqrt(1 - beta_a ** 2)
-        gamma_b = 1.0 / math.sqrt(1 - beta_b ** 2)
-        pA = math.sqrt(
-            (m_parent - m_vis_a - m_invis_a)
-            * (m_parent + m_vis_a - m_invis_a)
-            * (m_parent - m_vis_a + m_invis_a)
-            * (m_parent + m_vis_a + m_invis_a)
-        ) / (2 * m_parent)
-        pB = math.sqrt(
-            (m_parent - m_vis_b - m_invis_b)
-            * (m_parent + m_vis_b - m_invis_b)
-            * (m_parent - m_vis_b + m_invis_b)
-            * (m_parent + m_vis_b + m_invis_b)
-        ) / (2 * m_parent)
-        p_vis_a_boosted = gamma_a * (beta_a * math.sqrt(m_vis_a ** 2 + pA ** 2) + pA)
-        p_vis_b_boosted = gamma_b * (beta_b * math.sqrt(m_vis_b ** 2 + pB ** 2) + pB)
-        p_invis_a_boosted = gamma_a * (
-            beta_a * math.sqrt(m_invis_a ** 2 + pA ** 2) - pA
-        )
-        p_invis_b_boosted = gamma_b * (
-            beta_b * math.sqrt(m_invis_b ** 2 + pB ** 2) - pB
-        )
-        p_miss = p_invis_a_boosted + p_invis_b_boosted
-        theta = random.uniform(0, math.tau)
-        c = math.cos(theta)
-        s = math.sin(theta)
-        px_miss, py_miss = p_miss * c, p_miss * s
-        ax, ay = p_vis_a_boosted * c, p_vis_a_boosted * s
-        bx, by = p_vis_b_boosted * c, p_vis_b_boosted * s
+        chi_a = 0
+        chi_b = 0
 
         for mt2_impl in (mt2, mt2_arxiv):
-            val = mt2_impl(
-                m_vis_a, ax, ay, m_vis_b, bx, by, px_miss, py_miss, m_invis_a, m_invis_b
+            computed_val = mt2_impl(
+                m_vis_a, px_a, py_a, m_vis_b, px_b, py_b, px_miss, py_miss, chi_a, chi_b
             )
+            self.assertAlmostEqual(computed_val, 0.09719971)
+
+    def test_collinear_endpoint_cases(self):
+        random.seed(0)  # If the test fails, we want to be able to repeat the test!
+        for i in range(10000):
+            m_vis_a = random.uniform(0, 10)
+            m_vis_b = random.uniform(0, 10)
+            m_invis_a = random.uniform(0, 10)
+            m_invis_b = random.uniform(0, 10)
 
-            # passes with rel=1e-12 but sporadically fails with rel=1e-13
-            assert val == pytest.approx(m_parent, rel=1e-12), (
-                f"WARNING! Expected {m_parent} from collinear event but instead got "
-                f"{val} for mt2({m_vis_a},{ax},{ay}, {m_vis_b},{bx},{by}, "
-                f"{px_miss},{py_miss}, {m_invis_a},{m_invis_b}) in test case {i}."
+            # Addition of random positive number prevents possibility of division
+            # by zero at the locations below marked "MOO"
+            m_parent = max(m_vis_a + m_invis_a, m_vis_b + m_invis_b) + random.uniform(
+                0.1, 10
+            )
+            p_parent_a = random.uniform(0, 10)
+            p_parent_b = random.uniform(0, 10)
+            e_parent_a = math.sqrt(p_parent_a**2 + m_parent**2)
+            e_parent_b = math.sqrt(p_parent_b**2 + m_parent**2)
+            beta_a = p_parent_a / e_parent_a  # MOO
+            beta_b = p_parent_b / e_parent_b  # MOO
+            gamma_a = 1.0 / math.sqrt(1 - beta_a**2)
+            gamma_b = 1.0 / math.sqrt(1 - beta_b**2)
+            pA = math.sqrt(
+                (m_parent - m_vis_a - m_invis_a)
+                * (m_parent + m_vis_a - m_invis_a)
+                * (m_parent - m_vis_a + m_invis_a)
+                * (m_parent + m_vis_a + m_invis_a)
+            ) / (2 * m_parent)
+            pB = math.sqrt(
+                (m_parent - m_vis_b - m_invis_b)
+                * (m_parent + m_vis_b - m_invis_b)
+                * (m_parent - m_vis_b + m_invis_b)
+                * (m_parent + m_vis_b + m_invis_b)
+            ) / (2 * m_parent)
+            p_vis_a_boosted = gamma_a * (beta_a * math.sqrt(m_vis_a**2 + pA**2) + pA)
+            p_vis_b_boosted = gamma_b * (beta_b * math.sqrt(m_vis_b**2 + pB**2) + pB)
+            p_invis_a_boosted = gamma_a * (
+                beta_a * math.sqrt(m_invis_a**2 + pA**2) - pA
             )
+            p_invis_b_boosted = gamma_b * (
+                beta_b * math.sqrt(m_invis_b**2 + pB**2) - pB
+            )
+            p_miss = p_invis_a_boosted + p_invis_b_boosted
+            theta = random.uniform(0, math.tau)
+            c = math.cos(theta)
+            s = math.sin(theta)
+            px_miss, py_miss = p_miss * c, p_miss * s
+            ax, ay = p_vis_a_boosted * c, p_vis_a_boosted * s
+            bx, by = p_vis_b_boosted * c, p_vis_b_boosted * s
+
+            for mt2_impl in (mt2, mt2_arxiv):
+                val = mt2_impl(
+                    m_vis_a,
+                    ax,
+                    ay,
+                    m_vis_b,
+                    bx,
+                    by,
+                    px_miss,
+                    py_miss,
+                    m_invis_a,
+                    m_invis_b,
+                )
+
+                # passes with rel=1e-12 but sporadically fails with rel=1e-13
+                # update 2024-08-06: fails on MacOS unless tolerance doubled to 2e-12.
+                numpy.testing.assert_allclose(
+                    val,
+                    m_parent,
+                    rtol=2e-12,
+                    err_msg=(
+                        f"WARNING! Expected {m_parent} from collinear event but instead got "
+                        f"{val} for mt2({m_vis_a},{ax},{ay}, {m_vis_b},{bx},{by}, "
+                        f"{px_miss},{py_miss}, {m_invis_a},{m_invis_b}) in test case {i}."
+                    ),
+                )
diff --git a/tests/test_mt2_lally.py b/tests/test_mt2_lally.py
index 27b991b..6f92e19 100644
--- a/tests/test_mt2_lally.py
+++ b/tests/test_mt2_lally.py
@@ -1,74 +1,74 @@
 """Tests for the variant of MT2 by Colin Lally."""
 
+import unittest
+
 import numpy
-import pytest
-
-from .common import mt2_lally, mt2_lester
-
-
-def test_simple_example():
-    computed_val = mt2_lally(100, 410, 20, 150, -210, -300, -200, 280, 100, 100)
-    assert computed_val == pytest.approx(412.628)
-
-
-def test_near_massless():
-    # This test is based on Fig 5 of https://arxiv.org/pdf/1411.4312.pdf
-    m_vis_a = 0
-    px_a = -42.017340486
-    py_a = -146.365340528
-
-    m_vis_b = 0.087252259
-    px_b = -9.625614206
-    py_b = 145.757295514
-
-    px_miss = -16.692279406
-    py_miss = -14.730240471
-
-    chi_a = 0
-    chi_b = 0
-
-    computed_val = mt2_lally(
-        m_vis_a, px_a, py_a, m_vis_b, px_b, py_b, px_miss, py_miss, chi_a, chi_b
-    )
-    assert computed_val == pytest.approx(0.09719971)
-
-
-@pytest.mark.skip(reason="Currently failing due to inconsistencies")
-def test_fuzz():
-    batch_size = 100
-    num_tests = 1000
-
-    numpy.random.seed(42)
-
-    def _random_batch(min_, max_):
-        return numpy.random.uniform(min_, max_, (batch_size,))
-
-    for _ in range(num_tests):
-        m_vis_1 = _random_batch(0, 100)
-        px_vis_1 = _random_batch(-100, 100)
-        py_vis_1 = _random_batch(-100, 100)
-        m_vis_2 = _random_batch(0, 100)
-        px_vis_2 = _random_batch(-100, 100)
-        py_vis_2 = _random_batch(-100, 100)
-        px_miss = _random_batch(-100, 100)
-        py_miss = _random_batch(-100, 100)
-        m_invis_1 = _random_batch(0, 100)
-        m_invis_2 = _random_batch(0, 100)
-
-        args = (
-            m_vis_1,
-            px_vis_1,
-            py_vis_1,
-            m_vis_2,
-            px_vis_2,
-            py_vis_2,
-            px_miss,
-            py_miss,
-            m_invis_1,
-            m_invis_2,
-        )
 
-        result_lester = mt2_lester(*args)
-        result_lally = mt2_lally(*args)
+from tests.common import mt2_lally, mt2_lester
+
+
+class TestLally(unittest.TestCase):
+    def test_simple_example(self):
+        computed_val = mt2_lally(100, 410, 20, 150, -210, -300, -200, 280, 100, 100)
+        self.assertAlmostEqual(computed_val, 412.627668458219)
+
+    def test_near_massless(self):
+        # This test is based on Fig 5 of https://arxiv.org/pdf/1411.4312.pdf
+        m_vis_a = 0
+        px_a = -42.017340486
+        py_a = -146.365340528
 
-        numpy.testing.assert_allclose(result_lester, result_lally, rtol=1e-12)
+        m_vis_b = 0.087252259
+        px_b = -9.625614206
+        py_b = 145.757295514
+
+        px_miss = -16.692279406
+        py_miss = -14.730240471
+
+        chi_a = 0
+        chi_b = 0
+
+        computed_val = mt2_lally(
+            m_vis_a, px_a, py_a, m_vis_b, px_b, py_b, px_miss, py_miss, chi_a, chi_b
+        )
+        self.assertAlmostEqual(computed_val, 0.09719971)
+
+    @unittest.skip(reason="Currently failing due to inconsistencies")
+    def test_fuzz(self):
+        batch_size = 100
+        num_tests = 1000
+
+        numpy.random.seed(42)
+
+        def _random_batch(min_, max_):
+            return numpy.random.uniform(min_, max_, (batch_size,))
+
+        for _ in range(num_tests):
+            m_vis_1 = _random_batch(0, 100)
+            px_vis_1 = _random_batch(-100, 100)
+            py_vis_1 = _random_batch(-100, 100)
+            m_vis_2 = _random_batch(0, 100)
+            px_vis_2 = _random_batch(-100, 100)
+            py_vis_2 = _random_batch(-100, 100)
+            px_miss = _random_batch(-100, 100)
+            py_miss = _random_batch(-100, 100)
+            m_invis_1 = _random_batch(0, 100)
+            m_invis_2 = _random_batch(0, 100)
+
+            args = (
+                m_vis_1,
+                px_vis_1,
+                py_vis_1,
+                m_vis_2,
+                px_vis_2,
+                py_vis_2,
+                px_miss,
+                py_miss,
+                m_invis_1,
+                m_invis_2,
+            )
+
+            result_lester = mt2_lester(*args)
+            result_lally = mt2_lally(*args)
+
+            numpy.testing.assert_allclose(result_lester, result_lally, rtol=1e-12)
diff --git a/tests/test_mt2_tombs.py b/tests/test_mt2_tombs.py
index 5fe1933..1806f7d 100644
--- a/tests/test_mt2_tombs.py
+++ b/tests/test_mt2_tombs.py
@@ -1,97 +1,93 @@
 """Tests for the variant of MT2 by Rupert Tombs."""
 
-from typing import Optional, Union
+import unittest
 
 import numpy
-import pytest
-
-from .common import mt2_lester, mt2_tombs
-
-
-def test_simple_example():
-    computed_val = mt2_tombs(100, 410, 20, 150, -210, -300, -200, 280, 100, 100)
-    assert computed_val == pytest.approx(412.628)
-
-
-def test_near_massless():
-    # This test is based on Fig 5 of https://arxiv.org/pdf/1411.4312.pdf
-    m_vis_a = 0
-    px_a = -42.017340486
-    py_a = -146.365340528
-
-    m_vis_b = 0.087252259
-    px_b = -9.625614206
-    py_b = 145.757295514
-
-    px_miss = -16.692279406
-    py_miss = -14.730240471
-
-    chi_a = 0
-    chi_b = 0
-
-    computed_val = mt2_tombs(
-        m_vis_a, px_a, py_a, m_vis_b, px_b, py_b, px_miss, py_miss, chi_a, chi_b
-    )
-    assert computed_val == pytest.approx(0.09719971)
-
-
-def test_fuzz():
-    batch_size = 100
-    num_tests = 1000
-
-    numpy.random.seed(42)
-
-    def _random_batch(min_, max_):
-        return numpy.random.uniform(min_, max_, (batch_size,))
-
-    for _ in range(num_tests):
-        m_vis_1 = _random_batch(0, 100)
-        px_vis_1 = _random_batch(-100, 100)
-        py_vis_1 = _random_batch(-100, 100)
-        m_vis_2 = _random_batch(0, 100)
-        px_vis_2 = _random_batch(-100, 100)
-        py_vis_2 = _random_batch(-100, 100)
-        px_miss = _random_batch(-100, 100)
-        py_miss = _random_batch(-100, 100)
-        m_invis_1 = _random_batch(0, 100)
-        m_invis_2 = _random_batch(0, 100)
-
-        args = (
-            m_vis_1,
-            px_vis_1,
-            py_vis_1,
-            m_vis_2,
-            px_vis_2,
-            py_vis_2,
-            px_miss,
-            py_miss,
-            m_invis_1,
-            m_invis_2,
-        )
 
-        result_lester = mt2_lester(*args)
-        result_tombs = mt2_tombs(*args)
+from tests.common import mt2_lester, mt2_tombs
+
 
-        numpy.testing.assert_allclose(result_lester, result_tombs, rtol=1e-12)
+class TestTombs(unittest.TestCase):
+    def test_simple_example(self):
+        computed_val = mt2_tombs(100, 410, 20, 150, -210, -300, -200, 280, 100, 100)
+        self.assertAlmostEqual(computed_val, 412.627668458219)
 
+    def test_near_massless(self):
+        # This test is based on Fig 5 of https://arxiv.org/pdf/1411.4312.pdf
+        m_vis_a = 0
+        px_a = -42.017340486
+        py_a = -146.365340528
 
-def test_scale_invariance():
-    example_args = numpy.array((100, 410, 20, 150, -210, -300, -200, 280, 100, 100))
-    example_val = mt2_tombs(*example_args)
+        m_vis_b = 0.087252259
+        px_b = -9.625614206
+        py_b = 145.757295514
 
-    # mt2 scales with its arguments; check over some orders of magnitude.
-    for i in range(-100, 100, 10):
-        scale = 10.0 ** i
-        with numpy.errstate(over="ignore"):
-            # Suppress overflow warnings when performing the evaluation; we're happy
-            # so long as we match approximately in the test below.
-            computed_val = mt2_tombs(*(example_args * scale))
-        assert computed_val == pytest.approx(example_val * scale)
+        px_miss = -16.692279406
+        py_miss = -14.730240471
 
+        chi_a = 0
+        chi_b = 0
 
-def test_negative_masses():
-    # Any negative mass is unphysical.
-    # These arguments use negative masses to make both initial bounds negative.
-    # Check that the result is neither positive nor an infinite loop.
-    computed_val = mt2_tombs(1, 2, 3, 4, 5, 6, 7, 8, -90, -100)
-    assert not (computed_val > 0)
+        computed_val = mt2_tombs(
+            m_vis_a, px_a, py_a, m_vis_b, px_b, py_b, px_miss, py_miss, chi_a, chi_b
+        )
+        self.assertAlmostEqual(computed_val, 0.09719971)
+
+    def test_fuzz(self):
+        batch_size = 100
+        num_tests = 1000
+
+        numpy.random.seed(42)
+
+        def _random_batch(min_, max_):
+            return numpy.random.uniform(min_, max_, (batch_size,))
+
+        for _ in range(num_tests):
+            m_vis_1 = _random_batch(0, 100)
+            px_vis_1 = _random_batch(-100, 100)
+            py_vis_1 = _random_batch(-100, 100)
+            m_vis_2 = _random_batch(0, 100)
+            px_vis_2 = _random_batch(-100, 100)
+            py_vis_2 = _random_batch(-100, 100)
+            px_miss = _random_batch(-100, 100)
+            py_miss = _random_batch(-100, 100)
+            m_invis_1 = _random_batch(0, 100)
+            m_invis_2 = _random_batch(0, 100)
+
+            args = (
+                m_vis_1,
+                px_vis_1,
+                py_vis_1,
+                m_vis_2,
+                px_vis_2,
+                py_vis_2,
+                px_miss,
+                py_miss,
+                m_invis_1,
+                m_invis_2,
+            )
+
+            result_lester = mt2_lester(*args)
+            result_tombs = mt2_tombs(*args)
+
+            numpy.testing.assert_allclose(result_lester, result_tombs, rtol=1e-12)
+
+    def test_scale_invariance(self):
+        example_args = numpy.array((100, 410, 20, 150, -210, -300, -200, 280, 100, 100))
+        example_val = mt2_tombs(*example_args)
+
+        # mt2 scales with its arguments; check over some orders of magnitude.
+        for i in range(-100, 100, 10):
+            scale = 10.0**i
+            with numpy.errstate(over="ignore"):
+                # Suppress overflow warnings when performing the evaluation; we're happy
+                # so long as we match approximately in the test below.
+                computed_val = mt2_tombs(*(example_args * scale))
+            numpy.testing.assert_allclose(computed_val, example_val * scale)
+
+    def test_negative_masses(self):
+        # Any negative mass is unphysical.
+        # These arguments use negative masses to make both initial bounds negative.
+        # Check that the result is neither positive nor an infinite loop.
+        computed_val = mt2_tombs(1, 2, 3, 4, 5, 6, 7, 8, -90, -100)
+        self.assertLessEqual(computed_val, 0)
diff --git a/tox.ini b/tox.ini
deleted file mode 100644
index 6683f43..0000000
--- a/tox.ini
+++ /dev/null
@@ -1,41 +0,0 @@
-[tox]
-envlist = py39, py310, py311, py312, flake8
-
-[gh-actions]
-python =
-    3.9: py39
-    3.10: py310
-    3.11: py311
-    3.12: py312
-
-[testenv:flake8]
-basepython = python
-deps = flake8
-commands = flake8 mt2 tests
-
-[testenv:latestnpy]
-setenv =
-    PYTHONPATH = {toxinidir}
-deps =
-    -r{toxinidir}/requirements_dev.txt
-; If you want to make tox run the tests with the same versions, create a
-; requirements.txt with the pinned versions and uncomment the following line:
-;     -r{toxinidir}/requirements.txt
-commands =
-    pip install -U pip
-    pytest --basetemp={envtmpdir}
-
-[testenv:oldestnpy]
-setenv =
-    PYTHONPATH = {toxinidir}
-deps =
-    -r{toxinidir}/requirements_dev.txt
-; If you want to make tox run the tests with the same versions, create a
-; requirements.txt with the pinned versions and uncomment the following line:
-;     -r{toxinidir}/requirements.txt
-commands =
-    pip uninstall -y numpy
-    pip install -U pip
-    pip install oldest-supported-numpy
-    pytest --basetemp={envtmpdir}
-