Inspection guide · Updated May 6, 2026
NYC Restaurant Inspection Grades, Honestly Explained
There are two kinds of New Yorkers who know what an inspection grade card means. The first kind has run a kitchen. The second kind has gotten food poisoning at one. Everyone else has glanced at the colored letter on their way through the door and assumed someone else was paying attention. Someone is. NYC restaurant inspection grades are produced by a system that is rigorous, opinionated, and frequently misunderstood. Here is what those letters are actually telling you, by someone who has stood on both sides of the inspector's clipboard.
How NYC restaurant inspection grades are scored
An inspector from the Department of Health and Mental Hygiene shows up unannounced. They walk every station, take every temperature, look at every cooler, ask whoever is closest to wash their hands so they can watch how it's done. Each violation observed is point-valued against a fifty-three-item code. The score determines the letter:
- A: 0–13 points. The cleanest tier. Roughly four out of five inspections land here on a first visit.
- B: 14–27 points. More than one bad call, less than catastrophic.
- C: 28+ points. Significant. Triggers mandatory follow-up and, in some cases, closure.
- P (or Z) — Pending: The kitchen scored a B or C on the first visit and is awaiting re-inspection, which decides the final grade.
- N — Not yet graded: Usually a new establishment. Read it as "no signal yet," not as a problem.
The grade card collapses a lot of texture into a single character. That's the point of them — fast, public, comparable. But the texture matters once you know it's there.
What an inspector is actually looking at
Two categories of violation, with very different weights.
Critical violations are the ones a chef worth their salt understands viscerally: food held warm that should be cold, or cold that should be warm. Bare-hand contact with ready-to-eat product. An employee who ate a French fry, picked their nose, and went back to plating. Raw chicken juice trickling down through the cooler onto pre-portioned salad mix. These are not aesthetic offenses. They are how people get hospitalized. Each one is worth five to seven points, sometimes more if repeated.
Non-critical violations are the boring sins of a tired kitchen: the unlabeled prep, the missing thermometer, the can opener that hasn't been replaced since the second Bush administration. They are facility hygiene rather than food safety. Two to four points each. They add up.
Beyond those, there's a third category — public health hazards — which is what gets a kitchen shut down on the spot. An active rodent infestation. A backed-up grease trap leaking sewage onto the prep floor. Hot water out of service for the hand sink. Letters collapse all of this into one symbol, but the underlying record is itemized, and you can read it on the city's open data portal at data.cityofnewyork.us.
Why most B grades vanish
Most diners assume a B means the restaurant is now stuck with a B. That's not how the system works. After a B or C, the kitchen has thirty days to call the city back, fix what was cited, and earn an A on re-inspection. If they pass, the visible grade in the window resets to A. The original B never gets posted publicly.
This is also why so many restaurants have a perfect-looking grade in the storefront window but show a recent B in their public record. The grade card is the latest verdict. The history is the longer story. Both are useful.
What the system is rewarding here is improvement. A kitchen that takes a B, fixes the underlying problems, and earns its A back has demonstrated more than a kitchen that has never been challenged. We surface this as a "Violations Closed" badge in the directory, because the diner who only reads the storefront letter is missing the most interesting signal in the data.
What the letter card won't tell you
- Which violations were cited. A B-grade restaurant with one rodent observation is a different kind of problem than a B-grade restaurant with eight unlabeled prep containers. The score collapses both into the same letter. The detail is in the cited violations, which we display in full.
- When the inspection happened. An A from last week is a different signal than an A from eleven months ago. The kitchen of October is rarely the kitchen of March.
- The trajectory. A → A → A → B is one slip. A → B → A → B is a kitchen with a real management gap. Both can hold an A in the window. They are not the same restaurant.
- The closures. The DOHMH can shut a place down for a public health hazard. Closure orders show up in the inspection record, but not always on the visible grade card. They are the loudest signal in the data, and they are easy to miss if you only read the front window.
This is the gap the directory is meant to close. A 0–100 Clean Score that blends grade, violation count, recency, criticality, and trajectory into a single number you can rank by. Same data the city publishes. Better surface area.
How to use the system without overreacting to it
The temptation, once you understand how the system works, is to overcorrect — to refuse to eat anywhere that's ever been B, to treat a "Pending" sign as a verdict, to assume a perfect A is an absolute guarantee. None of these is right. The system is probabilistic. Use it that way.
- Start with A-only as a filter. Roughly four-fifths of NYC restaurants will pass it. The bar is low; the signal is mostly in what fails.
- Look at the inspection date. Anything older than nine months is a stale read.
- Read the latest inspection's actual violations. If the only thing cited was non-critical and the place fixed it on the spot, that's noise. If pest evidence is on the list, slow down.
- Look at history, not just the latest letter. A two-year run of A grades with zero criticals is the strongest signal the data offers. We sort by it.
Four myths that need to die
"An A means the food is safe." An A means thirteen or fewer points on one inspection on one day. The kitchen that earned it can still serve a bad oyster the next morning. Probabilistically, A-grade kitchens cause fewer foodborne illnesses, full stop. But "fewer" is not "none."
"A C means you'll get sick." Most C-grade kitchens don't make anyone sick on any given day. The grade is a risk signal, not a guarantee of harm. The reason the city is aggressive about C grades is that the statistical correlation between C and outbreaks is real and measurable. The grade is a blunt instrument calibrated against a public health problem.
"Big chains are always cleaner." Standardization helps, but a chain location with weak management can absolutely score a C. The grade is per-location, never per-brand. A Starbucks in midtown and a Starbucks in a forgotten corner of Queens are two different operations with two different inspection records. Treat them that way.
"The grade in the window is current." By law it should be. Enforcement is uneven. Always check the actual most-recent inspection date. Six months is the comfortable upper bound on how stale a window grade can usefully be.
The chef's read
Inspections are not the enemy of a good restaurant. They are an audit. A serious operator runs the same checks on their own kitchen every week — temps, hand-washes, pest evidence, dating — and treats the city's visit as a formality. The kitchens that struggle are the ones who only think about sanitation when the inspector is in the building. NYC restaurant inspection grades reward the first kind and surface the second. Use the data accordingly. The full inspection database is public at NYC Open Data and reflects every inspection back to 2011.
Eat clean. Eat safe. Eat well.