Apple, Google, and Mozilla are collaborating on the development of Speedometer 3, a browser benchmark that measures responsiveness
Context & Ripple Effects
Speedometer has been a contested yardstick: WebKit owned Speedometer 2.0, and Google used it in March 2022 to claim Chrome 99 on macOS beat Safari by 15% — a score Apple could have dismissed as a rival's test. With Speedometer 3, Apple, Google, and Mozilla are building the responsiveness benchmark together, so no single engine vendor controls the ruler.
This follows the same cross-vendor playbook as Interop 2022, where the four browser makers agreed to fix compatibility pain points developers flagged across 15 areas. The difference is scope: Interop targets what works identically everywhere, Speedometer 3 targets how fast it feels — and the collaboration eventually shipped as Speedometer 3.0, framed as creating a shared understanding of web performance.
First-order effects
- Each of the three engine teams now has to optimize against a benchmark it helped design, so future responsiveness claims — like Google's earlier 15-point-margin boasts — get made and judged on jointly governed numbers.
- Mozilla gains standing it lacked when its own Project Quantum speed push competed against WebKit's home-turf benchmark: Firefox's responsiveness work is now measured on a test it co-authored.
Second-order effects
- Benchmark-driven engineering priorities shift toward real-world responsiveness — startup, interaction latency, app-like behavior — the areas where Firefox claimed M1 wins in 2020 and Chrome claimed macOS wins in 2022, pulling marketing comparisons onto common ground.
- A jointly maintained benchmark raises the cost of gaming the test: any vendor optimizing narrowly for Speedometer 3 risks scrutiny from the co-authors who wrote it.
Third-order effects
- If the Interop-plus-Speedometer pattern holds, browser competition consolidates around jointly governed standards for both compatibility and performance, shrinking the space where a single vendor's benchmark can define the narrative.
- Benchmarks harden into market interfaces: procurement, reviews, and developer tooling increasingly cite co-governed scores, making measured responsiveness — not vendor marketing — the durable basis for engine comparison.
The trend: Browser makers are replacing vendor-owned benchmarks with jointly governed ones, extending the Interop model from compatibility into performance measurement.