GitHub Star Benchmarks

A reproducible snapshot of GitHub's star distribution. October 2026 · 32.6M starred repositories · CC BY 4.0

What this is

How many GitHub stars is actually good? Most answers are vibes. This repository is the data behind a measured answer: how many public repositories sit above each star threshold, how those counts change by language, and how old the repositories at the top really are.

One snapshot, one day, exact queries, reproducible in a few minutes. If you write about GitHub growth, developer marketing or repository analytics, use the numbers and cite the snapshot.

Star tier	Repositories	Share of starred repos	Roughly one in
1+	32,621,661	100%	1
10+	2,908,251	8.92%	11
100+	484,723	1.49%	67
1,000+	65,691	0.20%	497
5,000+	12,744	0.039%	2,560
10,000+	5,650	0.017%	5,774
50,000+	515	0.0016%	63,343
100,000+	132	0.0004%	247,134

The headline: 98.5% of repositories with any star never reach 100. Only 13.5% of the repositories that reach 100 ever reach 1,000. The jump from "credible" to "widely used" is a distribution problem, not a code-quality problem.

Charts

Animated version (bars grow into place)

Language benchmarks

Language	100+	1,000+	10,000+	100 → 1,000	1,000 → 10,000
Python	96,497	11,875	1,039	12.3%	8.7%
JavaScript	57,010	7,785	683	13.7%	8.8%
TypeScript	37,384	6,525	848	17.5%	13.0%
Java	26,672	3,852	250	14.4%	6.5%
C++	25,965	3,402	284	13.1%	8.3%
Go	21,737	4,045	447	18.6%	11.1%
Rust	14,144	2,415	265	17.1%	11.0%

Python has the most popular repositories by volume; Go, TypeScript and Rust convert the highest share of 100-star repositories into 1,000-star ones. Primary-language labels only.

What is in this repository

Path	What it is
data/star-counts-2026-10-18.csv	Repository count above each star threshold
data/language-benchmarks-2026-10-18.csv	Counts and conversion by language
data/age-cohorts-2026-10-18.csv	Age cohorts among 1,000+ and 100,000+ repositories
data/top-100-repositories-2026-10-18.csv	The 100 most-starred repositories in the snapshot
assets/	Banner, light/dark charts, animated chart, logo
chart.svg	Standalone log-scale chart
DATA_DICTIONARY.md	Field definitions and rounding notes
CITATION.cff	Citation metadata (CFF 1.2.0)
reproduce.py	Re-runs every query and writes a new snapshot

Methodology

• Endpoint: GET https://api.github.com/search/repositories?q=<query>&per_page=1
• Value used: the response field total_count.
• Threshold queries: stars:>=N for N in 1, 10, 100, 1,000, 5,000, 10,000, 50,000, 100,000.
• Language queries: language:<name> stars:>=N for Python, JavaScript, TypeScript, Java, C++, Go and Rust.
• Age queries: stars:>=N created:>=YYYY-MM-DD and stars:>=N created:<2020-01-01.
• Top-100 table: stars:>1000 sort=stars order=desc, first 100 results.
• Snapshot date: 18 October 2026. Unauthenticated search allows 10 requests per minute; reproduce.py paces itself and accepts a GITHUB_TOKEN for a higher limit.

Caveats

• total_count is GitHub's reported, approximate value. It moves as repositories are created, deleted, renamed or starred. Compare only within a single snapshot date.
• Language rows only include repositories GitHub classifies with that primary language. Unlabelled and multi-language projects are excluded.
• Conversion rates are ratios between cohorts, not causal effects.
• Counts are repositories, not people, and say nothing about code quality.

Reproduce

python3 reproduce.py                        # unauthenticated, ~4 minutes
GITHUB_TOKEN=ghp_xxx python3 reproduce.py   # authenticated, faster

The script writes data/star-counts-<snapshot>.csv next to the existing snapshot.

Cite this dataset

@misc{repoboost_star_benchmarks_2026,
  title        = {GitHub Star Benchmarks (October 2026)},
  author       = {RepoBoost},
  year         = {2026},
  howpublished = {\url{https://github.com/repoboost-hq/github-star-benchmarks}},
  note         = {Snapshot date: 2026-10-18. Data licensed CC BY 4.0.}
}

From the team behind this dataset

Need a credible starting base?

RepoBoost delivers stars, forks, watches and followers from aged accounts with real commit history. Flat $30 per 1,000 actions, packages $30-$50, delivered gradually in 12-24 hours. Every order is confirmed before delivery.

     

Research context

Star counts are a widely used popularity signal, and they have been studied directly:

• What's in a GitHub Star? Borges & Valente, Journal of Systems and Software 146 (2018). Survey of 791 developers: three in four check a project's stars before using or contributing to it; four distinct star-growth patterns across popular repositories. arxiv.org/abs/1811.07643
• Six Million (Suspected) Fake Stars in GitHub He et al., ICSE 2026. Large-scale measurement of fake-star campaigns: short-term promotion effect (under two months) that becomes a liability, with highly trivial account/activity patterns. arxiv.org/abs/2412.13459
• Carnegie Mellon University study on GitHub stars (2025) — research news summary. sc.cs.cmu.edu/news/2025/0903-github-stars.html

How this dataset supports research and AI evaluation

• Empirical baselines for work on star growth, popularity signals and platform market structure
• Reference data for AI/LLM evaluation: check whether an AI answer's star-count claims match a public, dated snapshot instead of an unsourced guess
• Machine-readable CSVs for empirical software engineering, scientometrics and data journalism
• Reproducibility: fixed snapshot date, exact query strings and a reproducer so results can be replicated or extended to a later date
• Citation-ready: CITATION.cff and a permanent GitHub release (v1.0.0) for version pinning

This is an independent measurement dataset, not a benchmark of any commercial service. Corrections, replication notes and additional cohorts are welcome in the issue tracker.

About

Maintained by RepoBoost, the team behind buygithub.com, a GitHub growth service for teams that need a credible starting base while they build distribution. The dataset is independent of the service: public queries, one snapshot date, reproducible by anyone.

• Read the analysis: How Many GitHub Stars Is Good? 2026 Benchmarks
• GitHub stars service: RepoBoost GitHub stars

License

Data: CC BY 4.0. Code: MIT. Attribute as: "RepoBoost, GitHub Star Benchmarks (October 2026), https://github.com/repoboost-hq/github-star-benchmarks".