There is no single THC level that defines “under the influence” across the United States. Legal limits range from zero (any detectable amount) to 5 nanograms per milliliter (ng/mL) of blood, depending on where you live. The most common per se limit is 5 ng/mL, but twelve states don’t allow any THC in your system at all while driving.
Legal THC Limits by State
States handle THC and driving in four distinct ways, and the differences are dramatic. Five states, including Illinois, Montana, Nevada, Ohio, and Washington, set specific per se limits for delta-9-THC in blood, ranging from 2 to 5 ng/mL. If a blood test shows you’re at or above that number, you’re legally impaired, period. It works the same way as the 0.08% blood alcohol limit: the number alone is enough for a conviction regardless of how you appeared to be driving.
Twelve states take a stricter approach with zero-tolerance laws. Arizona, Delaware, Georgia, Indiana, Iowa, Michigan, Oklahoma, Pennsylvania, Rhode Island, South Dakota, Utah, and Wisconsin prohibit driving with any detectable THC or its metabolites in your system. Since THC metabolites can linger in blood for days or even weeks after use, you could test positive long after any impairing effects have worn off.
Colorado uses what’s called a “permissible inference” law. If your blood THC is 5 ng/mL or higher, a jury is allowed to presume you were impaired, but you can present evidence to argue otherwise. Other states rely on a broader “under the influence” standard, where prosecutors must show that THC actually affected your ability to drive, typically through officer observations and field sobriety tests rather than a specific blood number.
What the Science Says About Impairment
Blood THC levels between 2 and 5 ng/mL are associated with meaningful driving impairment, particularly in people who don’t use cannabis regularly. The risk rises with concentration. One large study found that drivers with blood THC at or above 5 ng/mL were 6.6 times more likely to be responsible for a crash compared to drug-free drivers. That crash risk is comparable to driving with a blood alcohol level of 0.15%, nearly twice the legal alcohol limit.
Even lower concentrations carry measurable risk. Researchers who grouped drivers by blood THC ranges found that adjusted odds of impairment increased steadily: roughly 1.5 times higher at 1 to 2 ng/mL, about 2.1 times higher at 3 to 4 ng/mL, and 2.1 times or more at 5 ng/mL and above. In highway driving experiments, higher THC doses caused progressively worse lane-keeping, the same kind of weaving that characterizes drunk driving.
The picture gets more complicated when alcohol enters the mix. Even small amounts of alcohol and cannabis together can amplify each other’s effects on driving ability. The combination is associated with more severe driving outcomes than either substance alone, which means a THC level that might produce only mild impairment on its own could become dangerous after a single drink.
Why Blood THC Levels Are Unreliable Markers
Unlike alcohol, where blood concentration tracks closely with impairment, THC doesn’t behave predictably in the body. Blood THC peaks within about ten minutes of smoking and drops rapidly as the compound is absorbed into fat tissue and the brain. This means a very high blood reading could reflect a moment right after use when effects are still building, while a lower reading a couple of hours later might coincide with the period of peak cognitive impairment.
Frequent cannabis users present an even bigger problem for blood-based limits. A study of regular users found that nearly a quarter had resting blood THC levels above 2 ng/mL even before using cannabis, and 5% exceeded 5 ng/mL at baseline. Forty-three percent had detectable THC above the zero-tolerance threshold of 0.5 ng/mL. These people would fail a blood test in a zero-tolerance state without having used cannabis that day.
Tolerance further muddies the waters. In a clinical trial where experienced users controlled their own intake, driving impairment could not be reliably predicted from blood THC concentration, the THC content of what they smoked, or their history of use. Frequent users with high blood THC levels sometimes performed no differently on driving tasks than occasional users with lower levels. This disconnect is a core reason why setting a universal “impairment number” for THC remains scientifically controversial.
How Officers Assess THC Impairment
Because blood draws take time and THC levels shift quickly, roadside assessment plays a major role in cannabis-related traffic stops. Officers use field sobriety tests originally designed for alcohol, and some of these do pick up THC impairment, though not as reliably.
In a randomized clinical trial, participants who had smoked THC were classified as impaired by trained officers 81% of the time, compared to 49% of people who smoked a placebo. The Walk and Turn and One Leg Stand tests were most effective at distinguishing THC users from non-users. The Finger to Nose test and a test checking whether your eyes can converge on a close object also showed significant differences. The Modified Romberg test, which involves standing still with eyes closed, was not sensitive to THC.
The catch: these tests had a specificity of only about 51% for THC exposure, meaning nearly half of sober people were incorrectly flagged as impaired. And when researchers checked whether test results actually predicted driving performance on a simulator, the specificity dropped to 36%. Field sobriety tests catch most impaired drivers but also flag many who aren’t impaired at all.
Oral Fluid Testing at the Roadside
Some jurisdictions now use saliva-based screening devices that can detect THC within minutes. These tests are reasonably good at identifying whether THC is present in your blood, with a sensitivity of about 79% and a specificity of 98%. In practical terms, if you test positive on a saliva swab, there’s a very high chance THC is actually in your blood. If you test negative, there’s roughly a one-in-five chance of a missed detection.
The limitation is precision. Oral fluid THC concentration only explains about 29% of the variation in blood THC levels, so a saliva test can confirm recent use but cannot accurately tell how much THC is in your blood. This makes saliva devices useful as a screening step but not as a standalone tool for proving you’re above a specific legal threshold. A blood draw typically follows a positive roadside screen.
Practical Implications for Users
If you live in a zero-tolerance state, any cannabis use in the days before driving could put you at legal risk, even if you feel completely sober. Regular users are especially vulnerable because THC accumulates in body fat and slowly re-enters the bloodstream. In per se states with a 5 ng/mL limit, most occasional users will drop below that threshold within a few hours of smoking, but heavy users may stay above it for much longer.
Medical marijuana patients face the same legal thresholds as recreational users in most states. Having a medical card does not provide a defense against per se or zero-tolerance DUI charges. A few states allow medical patients to argue they weren’t actually impaired, but the burden of proof falls on the driver.
The safest approach is straightforward: THC impairs driving most significantly in the first one to three hours after use, and combining any amount of cannabis with even one alcoholic drink substantially increases your risk. The legal limits, whether 2 ng/mL, 5 ng/mL, or zero, were set with the understanding that the science connecting a specific blood number to a specific level of impairment is imperfect. They exist as legal lines, not biological ones.