Crates / ahash
Rust crate · real downloads · data Oct 24, 2026
ahash downloads, and should you depend on it?
A non-cryptographic hash function using AES-NI for high performance crates.io · docs · repository
Widely used and steady.
ahash: 2,097 crates depend on it; no release in the last 90 days, so its count is all steady use.
How this verdict is decided (4 rules)
- Usage: crates.io reverse dependencies (crates that list it as a dependency): 10,000+ = foundational, 1,000+ = widely used, 100+ = established, under 100 = few direct dependents.
- Direction: the median quiet day in the last 30 days of the window against the first 30, read against the median tracked crate (the whole field moved +36% from late summer to October, so a raw trend says nothing). 10%+ ahead of the field = gaining; 10%+ behind = losing ground; otherwise steady. Steady with no release in 2+ years = quiet for years.
- Release spikes: the median release-window day against the quiet-day baseline: under 1.15x = no spike, 1.15x to 1.3x = small bumps, 1.3x or more = jumps on release days.
- Release recency: days since the newest version that was not yanked, from crates.io publish times.
release window (release day + next day)quiet dayquiet-day median · download the 89 days as CSV
The numbers in one paragraphahash got 173,949,828 crates.io downloads from Jul 27, 2026 to Oct 23, 2026 with no new version in that window, so every day is a quiet day: a median of 1,968,676 a day, about 13,780,732 a week. With no release spikes to remove, its whole count is the steady baseline.
What this means if you're choosing a dependency
- Well trodden. 2,097 crates depend on it, so common problems tend to have public answers and breaking changes get noticed fast.
- No release in over a year (last: 0.8.12, May 8, 2025). Fine for a finished crate; check open issues and the latest commit before relying on it.
- Still pre-1.0 (0.8.12). Under Cargo's semver rules a 0.x minor bump may break you, so pin the minor version (
ahash = "0.8") and read the changelog on upgrades. - Fit check: license MIT OR Apache-2.0, edition 2018, needs Rust 1.60.0 or newer (from the latest version's manifest).
- Its raw count is a fair read. Under 1% of its downloads in the window came from release windows, so the raw number and the baseline (1,968,676 a day) tell the same story.
Alternatives that do a similar job
Crates that share its crates.io category or keywords, most similar first. Each line is the crate's own description, then how it compares on the same data.
Comparisons people search for
Its place in the dependency graph
Built on ahash
1 of the 300 crates we track depend on it directly (biggest first).
ahash builds on
Tracked crates its latest version (0.8.12) depends on (optional ones marked).
What the chart says
- Weekend quiet days ran at 59% of weekday quiet days (1,249,220 vs 2,109,759). A dip that deep is the signature of work-week CI pipelines and developers building on weekdays.
- Biggest day: 3,290,372 on Oct 22, 2026, a quiet day with no release nearby.
Every number
| All-time downloads | 853,536,387 (rank #120 on crates.io by all-time downloads) |
| Downloads in the window | 173,949,828 (Jul 27, 2026 to Oct 23, 2026, 89 days) |
| Within 2 days of a release | 0 (0.0%) |
| Quiet-day median (baseline) | 1,968,676 a day; weekdays 2,109,759, weekends 1,249,220 |
| Release-adjusted downloads | 173,949,828 (release-window days capped at the baseline) |
| Baseline rank | #204 of the 300 crates we track |
| Quiet-day trend | +29% raw, -5% against the median tracked crate (+36%) |
| Releases in the window | 0 |
| Versions published, ever | 95 |
| Latest version | 0.8.12, published May 8, 2025 (license MIT OR Apache-2.0, edition 2018, MSRV 1.60.0) |
| First published | Feb 28, 2019 |
| Reverse dependencies | 2,097 |
How we compute this
crates.io counts every download, including CI runners, mirrors and bots that fetch each new version, so raw counts run high (why, in The Manual). We split the last 89 days into release windows (the day a version of ahash was published, plus the next day) and quiet days, and use the median quiet day as the baseline: an upper bound on steady use, never a head count. The method and data come from our study of 90 days of crates.io data from 69 crates, featured in This Week in Rust #674; check any crate live with the real-downloads checker.
Badge for your README
Maintainers can show the baseline in a README. It is optional, updates when we refresh the data, and links back here. Preview, then copy:
README, light
README, dark
Publishing a crate of your own? How to publish a Rust crate to crates.io, step by step.
Categories
Sources
- Daily downloads: crates.io API, /crates/ahash/downloads (last 90 days, all versions), fetched Oct 24, 2026.
- Versions and publish times: crates.io API, /crates/ahash/versions, same fetch.
- Reverse dependencies: crates.io reverse dependencies, same fetch. Dependency edges between tracked crates: each crate's default version, crates.io API.
- Fetched politely (one request every 2 seconds, identified User-Agent) per the crates.io data access policy. Data is crates.io's; the release-window split, baseline and verdict rules are ours, CC BY 4.0.