About RealHumanBenchmark
RealHumanBenchmark was engineered to solve a glaring problem in the cognitive testing ecosystem: legacy benchmark tools have degraded into slow, ad-choked platforms plagued by frame drops, input latency, and intrusive data tracking.
The Problem With Legacy Sites
Over the past decade, millions of gamers, athletes, and enthusiasts relied on early web benchmarks to test reaction time and memory. However, as those platforms commercialized, they became bloated with heavyweight banner ad networks, auto-playing video players, and cross-site tracking scripts.
In a test where single milliseconds distinguish average humans from elite esports athletes, background JavaScript garbage collection and layout reflows introduce devastating timer jitter and false latency.
Our Modern Solution
RealHumanBenchmark is an independent, non-commercial engineering project dedicated to providing an uncompromising, high-frequency testing environment for everyone.
We combine monotonic high-resolution hardware timers (performance.now()), zero-lag pointer listeners, high-refresh-rate V-Sync alignment (144Hz, 240Hz, 360Hz), and a clean dark interface with zero ads and zero tracking.
Precision Engineering Standards
How our engine achieves laboratory-grade consistency inside standard web browsers.
Replaced unreliable Date.now() clock drift with monotonic performance.now() timestamps.
Immediate event registration on hardware micro-switch depress without synthetic 300ms mobile touch delay.
State transitions are tightly coupled to the browser compositor through requestAnimationFrame.
Zero third-party telemetry scripts, zero ad exchanges, and zero background CPU contention.
The 8 Cognitive & Reflex Batteries
Reaction Time
250 msVisual stimulus travels from retina to visual cortex in ~20-40ms, and motor cortex sends signals to finger muscles in ~20ms. Most delay occurs during neural processing in the brain.
Sequence Memory
Level 8-9Measures working memory capacity and chunking capabilities in the prefrontal cortex. Average human working memory holds roughly 7 ± 2 items simultaneously (Miller’s Law).
Aim Trainer
450 msTests the Fitts’s Law principle in neuromuscular control: the time required to rapidly move to a target area is a function of the ratio between distance and target width.
Chimp Test
10 - 11In 2007, researchers demonstrated that juvenile chimpanzees (notably Ayumu) possess eidetic photographic working memory vastly superior to average human adults.
Verbal Memory
35 - 45 wordsMeasures episodic recognition memory and hippocampal indexing. It highlights the susceptibility to false recognition and cognitive interference as the list grows.
Number Memory
7 digitsMeasures phonological loop capacity in Baddeley’s working memory model. Digit span tests are standard components of clinical neuropsychological batteries (e.g. WAIS).
Visual Memory
Level 11Assesses the visuospatial sketchpad and spatial resolution in the parietal cortex. Tested extensively in cognitive aging and athletic performance assessments.
Typing Test
45 - 60 WPMEvaluates procedural muscle memory, cognitive pipelining (reading ahead while your fingers execute previous words), and error monitoring.
How Percentiles Are Calculated
Our percentile models are derived from standard normal distribution curves (Gaussian bell curves) parameterized across millions of documented human trials. Each score is mapped to a z-score relative to global population mean (μ) and standard deviation (σ).
Ready to Benchmark Your Reflexes?
Start with our flagship sub-millisecond reaction time benchmark, or explore all 8 batteries.