GPU & WebGL Test — Browser Graphics Benchmark
See what GPU your browser is using and how fast it can render. Quick read + small benchmark — no install.
See what GPU your browser is using and how fast it can render. Quick read + small benchmark — no install.
From WebGL_debug_renderer_info — may be masked in privacy-sensitive browsers.
10,000 triangles per frame for 5 seconds. Higher FPS = faster GPU.
Some browsers (Safari notably, plus newer Chrome/Firefox in private mode) mask the actual GPU string with a generic 'WebKit' or 'Mozilla' label to prevent fingerprinting. Your real GPU is doing the work — you just can't read its exact name from the browser. Check OS settings (System Information on macOS, dxdiag on Windows) for the unmasked name.
WebGL is the browser API that lets web pages use your GPU for 3D rendering and accelerated 2D. Modern web apps (Figma, Google Maps, Photoshop Web, AAA browser games) need it. Our benchmark draws 10,000 triangles per frame and measures frame rate — a fast GPU hits 144+ fps, an integrated laptop GPU around 60, and very old hardware below 30.
Common causes: (1) browser is using software rendering — Chrome's chrome://gpu page shows whether hardware acceleration is on; (2) thermal throttling — laptops drop GPU clock when hot; (3) external monitor over USB-C with bandwidth limits; (4) battery saver mode capping performance. Plug in, cool down, retest.
Not at all directly. Native benchmarks like 3DMark have full hardware access and run dozens of scenarios. We only test what's available through WebGL — a small slice of GPU capability. Use this for a quick comparison between machines or to confirm hardware acceleration works; use 3DMark for actual GPU performance ranking.
No. The GPU vendor + renderer string is already exposed to every site you visit (it's part of the WebGL spec). Browsers progressively mask this for fingerprinting protection. Our test only reads — doesn't write or send anything.
WebGL frame timing is non-deterministic. Other tabs, system processes, browser GC, and OS scheduler all introduce jitter. Run the test 3 times and take the highest number for your reliable peak FPS. Big drops between runs usually mean another process is stealing GPU time.