Python Performance Benchmark Suite
1k
Stars
204
Forks
71
Open issues
30
Contributors
AI Analysis
pyperformance is an official Python benchmark suite maintained by the Python project, providing real-world performance measurements across Python implementations. It focuses on whole-application benchmarks rather than synthetic tests, serving as the authoritative source for CPython performance tracking. This tool is essential for Python core developers, performance engineers, and CPython maintainers; it is not suited for PyPy benchmarking, which has its own dedicated benchmark suite.
Inferred from signals mentioned in the README (tests, CI, type safety) — not a review of the actual code.
AI's overall editorial judgment — not an average of the bars above, can weigh other factors too.
Official Python benchmark suite for performance measurement across implementations
pyperformance is the authoritative benchmark collection maintained by the Python core team, designed to measure real-world performance across CPython, PyPy, and other Python implementations. It focuses on whole-application benchmarks rather than synthetic microbenchmarks. Adoption appears concentrated among Python core developers, language implementers, and performance-conscious infrastructure teams, rather than general application developers.
Created in 2016 as a centralized benchmark repository for the Python ecosystem. It emerged from the need for standardized, implementation-agnostic performance measurement tools. The project is maintained under the official Python organization on GitHub, lending it authority and longevity.
Growth has been gradual and linear rather than explosive. Star acquisition has plateaued (0 new stars in last 7 days), suggesting stable but not accelerating adoption. The project appears to serve a stable, narrow audience: those explicitly responsible for Python language or runtime performance. Recent activity (last push 2026-07-01) confirms continued maintenance, but growth signals are modest.
Adoption not formally verified in README. However, the official Python organization stewardship and positioning as 'authoritative source of benchmarks' suggests use by Python core developers and CPython performance teams. The fact that PyPy maintains separate benchmarks (noted in README) indicates pyperformance is specifically CPython-focused and may be used by CPython maintainers. Concrete evidence of enterprise or third-party adoption is absent from available metadata.
Based on README, pyperformance appears to be a collection of benchmark scenarios and measurement harness. It likely includes utilities for running, collecting, and analyzing performance data across multiple Python implementations. Specific implementation details cannot be verified from README alone.
not documented in README
Repository shows active maintenance as of July 1, 2026 (5 days before evaluation date). GitHub Actions CI is operational. However, modest commit frequency and zero stars gained in the past week suggest low velocity rather than rapid development. Activity is consistent with a project in stable maintenance mode rather than active feature development.
ADOPT IF: you are a Python core contributor, language implementer, or infrastructure engineer responsible for tracking CPython performance across versions and commits. AVOID IF: you need cross-implementation benchmarking (PyPy benchmarks are separate) or general-purpose microbenchmarking integrated with pytest. MONITOR IF: you are building Python language tooling and want alignment with official performance baselines, or if adoption beyond core teams increases.
Independent dimensions
Mainstream potential
3/10
Technical importance
7/10
Adoption evidence
3/10
- Fragmentation: PyPy and other implementations maintain separate benchmarks, reducing value as a unified standard
- Narrow audience: Adoption appears limited to Python core team and specialist performance engineers; limited mainstream developer mindshare
- Maintenance burden: Real-world benchmarks require ongoing curation to remain representative; README does not document governance for benchmark acceptance
- Competition: pytest-benchmark has higher adoption and broader ecosystem integration, may be preferred by teams not specifically aligned with CPython
- Documentation gaps: README does not detail how to contribute benchmarks, governance model, or performance targets
pyperformance will likely remain a stable, niche tool used primarily by CPython maintainers and language implementers for performance tracking. Adoption outside core Python teams appears unlikely to accelerate significantly without broader marketing or integration with popular testing frameworks.
Newsletter
Get analyses like this every Monday
Free weekly digest of the most interesting open-source discoveries.
Languages
Information
- Language
- Python
- License
- MIT
- Last updated
- 1w ago
- Created
- 120mo ago
- Analyzed with
- anthropic/claude-haiku-4-5
Stars over time
Contributors over time
Top 100 contributors only — repos with more will plateau at 100.
Open issues
We should vendor the Pyston macrobenchmarks
Enable decimal benchmarks
No Windows `build.bat` support in `compile`
asyncio_websockets tests the implementation of `zlib`, not of Python
Add benchmarks for complex type
Top contributors
Recent releases
Similar repos
ionelmc/pytest-benchmark
pytest-benchmark is a pytest fixture that enables benchmarking of Python code...
| Repository | Stars | Week Δ | Language | Score | Updated |
|---|---|---|---|---|---|
|
python/pyperformance
this
|
1k | — | Python | 8/10 | 1w ago |
|
ionelmc/pytest-benchmark
|
1.4k | — | Python | 8/10 | 3w ago |
 pytorch/benchmark
|
1k | — | Python | 7/10 | 2d ago |
 pypy/pypy
|
1.8k | — | Python | 8/10 | 15h ago |
 phpbench/phpbench
|
2k | — | PHP | 8/10 | 4w ago |
 benfred/py-spy
|
15.3k | — | Rust | 8/10 | 3w ago |
More stars (1,439 vs 1,020) and broader ecosystem integration. pytest-benchmark is general-purpose and developer-friendly; pyperformance is authoritative but narrower in scope, focused on standard workloads rather than arbitrary code profiling.
PyPy maintains its own separate benchmark suite (noted in README). pyperformance explicitly states it is 'not tuned for PyPy yet', creating a fragmented benchmark landscape across Python implementations rather than unified measurement.
Profiler and sampling tool (15,313 stars) vs benchmark suite. Different use case: py-spy is for investigating existing code; pyperformance is for structured performance regression testing.
Domain-specific benchmarks for machine learning (1,039 stars). Similar adoption scale but serves a specialized vertical rather than general Python performance measurement.
