GoKawiilLibsodium benchmarks show WebAssembly runtimes closing gap with native code in 2026
A developer re-ran libsodium crypto benchmarks across native x86-64 code and several WebAssembly runtimes (Wasmer, WAVM, WAMR, Wasmtime, wasm2c) using builds from mid-2024, mid-2025 and mid-2026 to see whether WASM performance is genuinely improving over time. The results show Wasmer leading, with WAVM, WAMR and Wasmtime close behind, and WAVM's optimizer notably strong at speeding up baseline portable WASM without extra instruction extensions.
GoKawiil's interpretation of the reporting above, not reported fact.
These findings matter because they move beyond marketing claims about WebAssembly speed and instead track real year-over-year progress using consistent, reproducible crypto workloads. The emergence of new instructions like wide_arithmetic shows that runtime performance gains increasingly depend on both engine optimization work and evolving WebAssembly standards being adopted together.
- Wasmer currently performs best among tested WebAssembly runtimes, with WAVM, WAMR and Wasmtime close competitors.
- WAVM's optimizer is particularly effective at accelerating baseline, portable WebAssembly code.
- New wide_arithmetic instructions significantly boost performance for cryptographic workloads when runtimes support them.
Source: 00f.net — Frank Denis, 2026-09-09
Published there as: “Performance of WebAssembly Runtimes in 2026”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.