AS SSD Benchmark vs CrystalDiskMark: Which SSD Test Should You Trust?
If you search for an SSD speed test, two names come up again and again: AS SSD Benchmark and CrystalDiskMark. They are both free, both small, both widely quoted in reviews — and their numbers cannot be swapped for each other. This page explains what each tool actually measures, how their test items line up, why the same drive produces different-looking results in each one, and how reviewers combine the two instead of picking a winner. If you only need the software itself, you can download AS SSD Benchmark first and run both tools side by side as you read.
The short answer
Use CrystalDiskMark when you want a current, actively maintained, general-purpose speed check with configurable queues and threads. Use AS SSD Benchmark when you want its specific combination of 4K and 64-thread results rolled into a single total score, plus its Copy and Compression tests and its alignment and controller display. Neither result is “the true speed” of your drive. Each is a defined workload, and a result only means something when you compare it with results from the same tool, the same test size, and similar settings.
How the test items correspond
The two tools overlap more than their layouts suggest. AS SSD Benchmark runs sequential read and write, random 4K read and write, 4K with 64 threads, and access time, using a 1 GB test file by default (1, 3, 5 or 10 GB can be selected) and without using the operating system cache for its synthetic tests. CrystalDiskMark, which is built on Microsoft DiskSpd, runs sequential tests and random 4K tests with queue and thread combinations you can configure. Roughly speaking, AS SSD’s Seq row corresponds to CrystalDiskMark’s sequential rows, AS SSD’s 4K row corresponds to a single-queue, single-thread random test, and the 4K-64Thrd row is a heavily queued random test — conceptually closer to CrystalDiskMark’s high-queue random presets, but not identical in block handling, timing or data pattern.
AS SSD Benchmark also adds two things CrystalDiskMark does not focus on. The Copy-Benchmark times three realistic copy scenarios — ISO (large files), Program (many small files) and Game (a mix) — using the operating system’s copy functions. The Compression-Benchmark shows how performance changes as data becomes more or less compressible, which matters on controllers whose speed depends on the data pattern. At the top of its window, AS SSD also shows the drive model, firmware, controller driver and partition alignment status, which is why it has long been a quick diagnostic as well as a benchmark.
Why the scores cannot be compared directly
AS SSD Benchmark collapses its results into Read, Write and Total scores. The simplified scoring formula is: Read score = Seq Read × 0.1 + 4K Read + 4K-64Thrd Read, with the same pattern for Write. That weighting deliberately makes small random transfers dominate the headline number. CrystalDiskMark does not produce a single score at all — it reports MB/s (and optionally IOPS) per test. So an AS SSD total of, say, a few thousand points and a CrystalDiskMark sequential figure of several thousand MB/s are not two estimates of the same thing. They are different units answering different questions.
Even the shared measurements differ underneath. The tools use different test engines, different default file sizes and different queue behaviour, so the same drive can show a higher sequential number in one tool and a stronger random result in the other. That is not a contradiction and it does not mean one tool is lying. It means you should never rank two drives where one was measured with AS SSD and the other with CrystalDiskMark, and you should never paste a CrystalDiskMark result into a thread asking for AS SSD scores. Compare like with like, and state the tool, version, test size and drive state whenever you publish a result.
Side-by-side comparison
| Dimension | AS SSD Benchmark | CrystalDiskMark |
|---|---|---|
| What it measures | Seq, 4K, 4K-64Thrd, access time, plus Copy and Compression tests; shows alignment and controller status | Sequential and random 4K workloads based on Microsoft DiskSpd, with configurable queues and threads |
| Score / result form | MB/s or IOPS per row, plus Read, Write and Total scores using a simplified weighted formula | MB/s (and IOPS) per test; no single total score |
| Platform | Windows only; portable, no installation | Windows |
| Licence and price | Freeware by Alexej Schepeljanski; free to use | Free; MIT licence (per Wikipedia), by Noriyuki Miyazaki / Crystal Dew World |
| Maintenance status | Final version 2.0.7316.34247, January 2020; no newer release since | Actively maintained; version 9.0.3 released 24 May 2026 |
| Best for | A single comparable total score, copy and compression behaviour, alignment checks, matching the large archive of published AS SSD results | A quick, current, configurable speed check and comparing drives tested with the same preset |
How reviewers use both together
Review sites have paired these tools for years because they are strong in different places. CrystalDiskMark gives a fast, repeatable picture of peak sequential throughput and of random performance at chosen queue depths — useful for checking that a new drive reaches roughly the speeds its class should reach. AS SSD Benchmark adds the weighted total score that makes drives easy to rank within its own database of results, the Copy test that resembles moving real folders around, and the compression graph that can reveal controllers that slow down on incompressible data. Running both takes only a few minutes, and together they catch problems that either tool alone can miss — for example a drive with strong sequential numbers but weak 4K results, which will feel slow as a system drive despite an impressive headline figure.
A sensible routine for a new drive is: run CrystalDiskMark for a baseline, run AS SSD Benchmark for the total score and copy results, save a screenshot or export from each, and repeat the same pair after any firmware, driver or alignment change. If you are choosing a drive rather than testing one, read our overview of the best SSD benchmark tools for the wider toolkit, and see our notes on choosing an NVMe SSD for how benchmark results translate into a buying decision.
Verdict by scenario
Checking a brand-new drive quickly: CrystalDiskMark, because it is current and configurable. Comparing your result with years of published scores, or checking alignment: AS SSD Benchmark. Diagnosing why a system feels slow: run both — the 4K rows matter more than sequential peaks for everyday responsiveness. Publishing a result for others to verify: publish both, with tool versions and test sizes stated. There is no loser here; there are only results used inside or outside the context that makes them meaningful.
Ready to run the comparison yourself? Get the official file from our download page, and read the companion comparison AS SSD Benchmark vs CrystalDiskInfo to see why a benchmark still tells you nothing about drive health.