Good Little Tools
JavaScript / CPU work
Simulation, sorting, hashing, and array math on the main thread.
—Waiting
Browser-native performance lab
How game-ready is
this browser?
Test the browser work that web games actually depend on: script execution, 2D and 3D drawing, frame consistency, display cadence, and feature support. No download. No account. No invented hardware specs.
Keep this tab visible and close heavy apps for a cleaner comparison.
GLT browser score
—
Ready when you are
The composite summarizes only tests completed in this browser. It is normalized to a fixed reference workload, not a global device percentile.
Live test board
0 / 6 completeSix honest signals
Canvas 2D
Sprite-like transforms, fills, curves, and compositing.
—Waiting
WebGL graphics
Shader-driven triangles at a workload sized to this viewport.
—Waiting
Frame responsiveness
Animation cadence and long-frame pressure during mixed work.
—Waiting
Refresh-rate estimate
Observed requestAnimationFrame cadence while this tab is visible.
—Waiting
Site latency sample
Round trips to this site—not your ISP speed or game-server ping.
—Waiting
Compatibility scan
What this browser brings to the game
Support is not the same as speed. These checks tell you which browser APIs a web game may be able to use.
Result notes
Read the score like a grown-up.
A browser test is a controlled snapshot, not a spec sheet. Use it to compare two browsers on the same device, check a setting change, or spot a regression after an update.
Reduce the noise
Plug in laptops, turn off battery saver, close busy tabs, and run twice. Thermal state and background work matter.
Prioritize consistency
A stable frame cadence can feel better than a higher peak. The responsiveness result calls out missed and long frames.
Keep conditions matched
Use the same window size, display, power mode, browser version, and test page. Mobile and desktop scores are not directly equivalent.
Method & limitationsMeasured here.
Measured here.
Claimed carefully.
What the composite means
The GLT score combines completed JavaScript (30%), Canvas (15%), WebGL (25%), and responsiveness (30%) sub-scores. Each is mapped to a fixed reference workload and capped at 200. Refresh cadence, network latency, and feature support are reported separately because they describe the environment, not raw compute.
Why it is not a hardware benchmark
Browsers add their own compiler, graphics translation, security, power, and scheduling layers. The same hardware can score differently across browsers and settings. We intentionally avoid naming a CPU/GPU from browser-exposed strings or predicting native-game FPS.
About WebGPU
WebGPU support is detected and shown as a capability. This short benchmark uses WebGL for its broadly comparable graphics workload; a WebGPU-only score would exclude many otherwise capable browsers and would not be directly comparable.
About latency
The latency sample times a handful of cache-busted requests to a tiny file on goodlittletools.com. Routing and server location dominate it. It does not measure download bandwidth, loaded latency, Wi-Fi quality in isolation, or ping to a particular game server.
A little healthy skepticism
Can this tell me how many FPS a specific game will run?
No. A game's engine, scene complexity, resolution, server, and native graphics API all matter. This measures a small browser-native workload and reports its own observed frame behavior only.
Why did my score change?
Short tests are sensitive to background work, temperature, power saving, extensions, browser updates, and window size. Run twice under matched conditions and treat small differences cautiously.
Is a higher refresh estimate always better?
No. It estimates the browser callback cadence for the active display. It can be capped, affected by variable refresh, or rounded. It is not a panel certification.
Does anything leave my device?
All benchmark computation and scoring stays in this page. Only the clearly labeled latency sample makes small requests to this site's own origin. The benchmark does not submit or store your results.