Meat Cooking Temperature Converter (°C, °F, Doneness)

Internal temperatures for meat by food and by doneness, in °C and °F both ways, with the USDA FSIS safe minimums reproduced exactly and kept separate from the culinary doneness scale — and a carryover-cooking estimate that says what drives it and how uncertain it is.

Last updated 24 Sep 2026

Meat temperature and pull temperature

Food and doneness → the temperature to pull at

What are you starting from?

Which food? (this is the control that sets the safe minimum)The eight categories USDA FSIS publishes a minimum for. The figure for each is in the first table below.

How done do you want it?The culinary doneness scale, from Certified Angus Beef’s own chart. It describes beef, and the page applies it only to whole cuts of red meat. For poultry, ground meat, eggs, ham, fish and reheated dishes it is ignored and the FSIS minimum is used instead.

The temperature you haveType in either unit; pick the matching mode above and both appear.

What shape is the piece?Shape, not size. It decides how much heat is stored in the outer layers relative to the middle, and it turns out to matter far more than thickness does.

Surface temperature when it comes off the heatA wet roast surface sits near 100 °C however hot the oven is, because the water boiling there pins it. A dry seared crust or a grilled surface runs hotter. The fourth table gives figures.

How much of the stored heat reaches the middle?Two bodies of evidence disagree by about a factor of two here. The default is the published rule of thumb, which predicts the SMALLER rise and therefore the later pull — the safer error.

A flat piece of meat seen end-on, not a circuit. The curve is the temperature profile through it at the moment it leaves the heat: hot at both surfaces, coolest in the middle, and very nearly parabolic, which is what a quasi-steady conduction profile looks like. The area between that curve and the centre line is the stored heat that has nowhere to go but inwards, and the fraction of it that reaches the middle is n/(n+2) — one third for this flat shape, a half for a cylinder, three fifths for a ball. The proportions of the drawing are illustrative; the numbers beneath it are live.

59.1°CExample

A rolled beef roast, cooked to USDA’s 145 °F whole-cut minimum, surface at 100 °C, using the published rule of thumb

One temperature gap, one shape factor, one honest fudge

T_pull = T_target − c × n/(n+2) × (T_surface − T_target)

T target: the internal temperature you want when it reaches the plate — from the doneness scale, or the FSIS minimum, or typed in
T surface: how hot the outside is when it leaves the heat. About 100 °C while the surface is wet, whatever the oven is set to; 120 to 160 °C for a dry seared crust
n: 1 for a flat piece heated from two sides, 2 for a cylinder, 3 for a ball
n/(n+2): the fraction of the stored surface-to-centre excess that an equilibrating piece delivers to its own middle: 1/3, 1/2, 3/5. Derived by integrating the quasi-steady conduction profile over the volume
c: how much of that survives the rest instead of escaping from the surface. 0.20 from the published rules of thumb, 0.5 to 0.8 from a finite-difference solve of the heat equation. The page defaults to the small one

Worked example

A rolled beef roast, cooked to USDA's 145 °F whole-cut minimum, surface at 100 °C, using the published rule of thumb
The target is 145 °F, which is exactly 62.7778 °C — and which is both the USDA FSIS safe minimum for a whole cut AND what a kitchen calls medium
The surface is at 100 °C, so the stored gap is 100 − 62.78 = 37.22 °C
A cylinder has n = 2, so n/(n+2) = 0.5 of that gap would reach the centre if nothing escaped
At c = 0.20 the rise is 0.20 × 0.5 × 37.22 = 3.72 °C, which is 6.7 Fahrenheit degrees — inside Certified Angus Beef's own 5 to 10 °F for a roast
So pull at 62.78 − 3.72 = 59.1 °C — and then read the warning, because that is below the safe minimum and FSIS wants the meat to reach 145 °F on the heat

USDA FSIS safe minimum internal temperatures

These are the figures, and they are not this page’s opinion. FSIS publishes them in Fahrenheit and prints its own Celsius rounding beside each one; both columns are here because they differ. Note the third row of the ground-meat line: 160 °F is 71.1 °C, and the “71 °C” that half the internet prints is a tenth of a degree BELOW a legal minimum. This page carries 71.1. FSIS’s own wording is that the food must reach the listed temperature as measured with a food thermometer **before removing it from the heat source**, which is the sentence that decides what carryover cooking is allowed to be used for.
Food	Safe minimum, °F	Exact °C	FSIS prints	Rest	Note
Beef, pork, veal or lamb — steaks, chops and roasts (whole cuts)	145	62.8	62.8 °C	3 min	The only FSIS category with a rest time written into the minimum.
Ground meat — beef, pork, veal or lamb	160	71.1	71.1 °C	None	Grinding spreads surface bacteria through the whole mass, so there is no safe rare burger.
All poultry — whole bird, breasts, legs, thighs, wings, ground, and stuffing	165	73.9	73.9 °C	None	One figure for every part and every form, including the stuffing.
Ham, fresh or smoked, uncooked	145	62.8	62.8 °C	3 min	Same as any other whole pork cut, rest included — but the red-meat doneness scale does not belong on pork, so the page does not offer it.
Ham, fully cooked, being reheated	140	60.0	60 °C	None	140 °F only if it was packaged in a USDA-inspected plant. Everything else reheats to 165 °F.
Egg dishes	160	71.1	71.1 °C	None	Eggs cooked in the shell or fried go by sight: yolk and white both firm.
Fish and shellfish	145	62.8	62.8 °C	None	Or, FSIS allows, until the flesh is opaque and separates with a fork. Fish has its own doneness conventions and they are not the beef ones, so the page does not apply a scale to it.
Leftovers and casseroles	165	73.9	73.9 °C	None	Reheating is not cooking again from raw; the whole dish has to reach it.

The culinary doneness scale, and where it crosses the safe minimum

Two scales, one thermometer, and the row that matters is “medium”. USDA’s safe minimum for a whole cut of beef, pork, lamb or veal is 145 °F, and 145 °F is what a kitchen calls medium. Everything a steakhouse calls rare or medium rare is below the USDA minimum. That is not a secret and it is not a scandal: it is a risk that millions of people choose knowingly every day, and whole-muscle beef carries it on its outer surface rather than through the middle, which is why searing matters. What this page will not do is print one of those numbers as though it were the other. Doneness figures are Certified Angus Beef’s own published chart — except the last row. Their chart stops at medium well, 150 °F, and publishes no well-done temperature for a steak or a roast at all; the 160 °F on that row is USDA’s own figure, which USDA itself glosses as “well done” when it gives the ground-beef minimum. It is on the page because people ask for it, and it is labelled because it did not come from where the rows above it came from.
Doneness	°F	Certified Angus Beef prints	Exact °C	Against the 145 °F whole-cut minimum
The USDA safe minimum for this food	0	—	-17.8	Whatever FSIS publishes
Rare	125	52 °C	51.7	Below it by 11.1 °C — a risk you are choosing
Medium rare	135	57 °C	57.2	Below it by 5.6 °C — a risk you are choosing
Medium	145	63 °C	62.8	This IS the minimum
Medium well	150	66 °C	65.6	Above it by 2.8 °C
Well done	160	—	71.1	Above it by 8.3 °C

Carryover: the model, the rule of thumb, and the factor of two between them

All six figures are for a piece pulled at 145 °F with its surface at 100 °C. The factor n/(n+2) is not a fudge: it is the fraction of the stored surface-to-centre excess that an equilibrating piece delivers to its own middle, and it comes out of integrating the quasi-steady conduction profile over the volume — one third for a slab heated from two sides, a half for a cylinder, three fifths for a sphere. A finite-difference solve of the heat equation confirms it and puts c between 0.48 and 0.89 over 72 combinations of shape, thickness, oven temperature and rest-side cooling. The published rules of thumb (Certified Angus Beef: pull roasts 5 to 10 °F low, steaks 5 °F low) imply c near 0.20. The one published MEASUREMENT anybody has put a number on — a 1½ inch steak pulled at 130 °F and arriving at 145 °F — needs c = 0.55. The two disagree by about a factor of two, the page defaults to the conservative end, and the real answer is to leave the probe in.
Shape	n	n/(n+2)	Rise at c = 0.20 (rule of thumb)	Rise at c = 0.65 (conduction model)	In Fahrenheit degrees
Flat	1	0.333	2.5 °C	8.1 °C	4 to 15 °F
Long and round	2	0.500	3.7 °C	12.1 °C	7 to 22 °F
Ball-shaped	3	0.600	4.5 °C	14.5 °C	8 to 26 °F

Surface temperature at the moment you pull it, from the finite-difference solve

Computed, not measured: a one-dimensional solve of the heat equation in a flat piece, pulled when the centre reaches 145 °F, with the surface held at 100 °C for as long as it is wet. Read along a row and the point of the page appears — the surface saturates. Once it is boiling, a hotter oven cooks the meat faster but does not raise the carryover at all, because the thing that drives carryover is the surface-to-centre gap and that gap has a ceiling. Oven temperature only moves the answer below about 110 °C. A grill, a broiler or a hard dry sear breaks the ceiling, because a browned crust is dry and can sit at 120 to 160 °C — put that number in the box instead and the estimate rises with it.
Thickness or diameter	120 °C oven	160 °C oven	200 °C oven	230 °C oven
15 mm	65 °C	72 °C	78 °C	83 °C
25 mm	71 °C	80 °C	90 °C	97 °C
40 mm	77 °C	91 °C	100 °C	100 °C
60 mm	84 °C	100 °C	100 °C	100 °C
90 mm	91 °C	100 °C	100 °C	100 °C
130 mm	97 °C	100 °C	100 °C	100 °C

The safe minimum, the doneness you want, and the gap between them

There are two temperatures on this page and they are not the same number. One is the USDA safe minimum internal temperature: a published figure, the same for everybody, the temperature at which a food is considered safe. The other is the temperature a cook actually takes the meat off the heat at, which is lower, because the meat keeps cooking after it leaves the oven. Almost every chart on the internet prints one of those and calls it the other. This one gives both and labels them.

The safe minimums are USDA’s, not this page’s. Whole cuts of beef, pork, lamb and veal, 145 °F, 62.78 °C, with a three-minute rest. Ground meats, 160 °F, 71.11 °C. All poultry, every part and every form, 165 °F, 73.89 °C. They are reproduced here in full because USDA FSIS is a US Government body and its chart is a public-domain work, and they are reproduced exactly: the exact Celsius conversion sits beside FSIS’s own rounding in the first table, because they differ, and 160 °F is 71.1 °C rather than the 71 °C that gets printed everywhere. A tenth of a degree is not a meaningful amount of food safety, but rounding a legal minimum downwards is the wrong habit to be in.

The doneness scale is a different thing entirely. Rare, medium rare, medium, medium well: those are Certified Angus Beef’s published figures, they describe what the middle of a steak looks like, and they have nothing to do with safety. (Their chart stops there. It publishes no well-done temperature for a steak at all, and the 160 °F on the last row of the table below is USDA’s own ground-beef figure, which USDA glosses as well done — labelled, because it came from somewhere else.) The one place the two scales touch is worth pausing on. USDA’s minimum for a whole cut is 145 °F. A kitchen calls 145 °F medium. Everything below medium — every rare and every medium rare steak ever served — is below the USDA safe minimum. This page will say that and then give you the number anyway, because the risk on an intact muscle is on the surface and searing addresses the surface, and because an adult is entitled to choose it. What it will not do is offer you a rare burger: grinding moves the outside to the inside, which is the whole reason ground meat has a minimum 15 °F higher than the muscle it came from.

Carryover cooking, and why the usual rule is not quite a rule. When meat comes out of the oven its outside is much hotter than its middle, and that heat has nowhere to go but inwards. The centre keeps climbing for ten to thirty minutes. How far it climbs is set by one quantity — how much hotter the surface is than the centre — multiplied by two factors. The first is the shape, and it is exact: integrate the quasi-steady conduction profile over the volume and a flat piece delivers one third of the stored excess to its own middle, a cylinder one half, a ball three fifths. The second is how much of that escapes through the surface while the piece rests, which is where the honesty comes in. A finite-difference solve of the heat equation puts it between 0.48 and 0.89 across 72 cases. The published rules of thumb imply about 0.20. They disagree by a factor of two and this page carries both, defaulting to the conservative one, because pulling late costs you a little pink and pulling early costs you something worse.

What is not driving it. Weight, mostly. And oven temperature, above about 110 °C — which is the finding this page would most like to be remembered for. While the surface of a roast is wet it cannot get above 100 °C, because the water boiling there pins it, so a 230 °C oven and a 160 °C oven arrive at the same surface temperature and therefore the same carryover. A hot oven cooks faster; it does not overshoot harder. What does break the ceiling is a dry crust — a hard sear or a grill can put the surface at 120 to 160 °C — and the surface temperature is an input on this page precisely so that you can say so.

One thing carryover is never for. FSIS’s wording is that the food must reach the listed temperature before removing it from the heat source. So carryover is how you land on a doneness target above the minimum without sailing past it. It is not a way of pulling below the minimum and hoping. The page always shows the lowest pull temperature that still reaches the minimum on the heat, and when your target needs you to pull below that, it says so in red-letter terms rather than quietly rounding your answer up.

For the oven side of the same problem — gas marks, fan offsets and the fact that gas mark 4 is 350 °F and therefore 176.7 °C and not the 180 °C every chart prints — the oven temperature converter is the page. For scaling a roast up or down, the recipe scaling calculator handles the tin and the timing, and the salt conversion calculator is where the brining arithmetic lives.

Frequently asked questions

What is the USDA safe minimum internal temperature?

It depends on the food, and there are eight figures rather than one. Whole cuts of beef, pork, lamb and veal: 145 °F, 62.8 °C, with a three-minute rest. Ground meats: 160 °F, 71.1 °C. All poultry, in every form including ground and including the stuffing: 165 °F, 73.9 °C. Fish and shellfish: 145 °F. Egg dishes: 160 °F. Leftovers and casseroles: 165 °F. The full chart, with FSIS’s own Celsius rounding beside the exact conversion, is the first table on this page.

Is medium rare steak safe?

Medium rare is about 135 °F, 57.2 °C, and the USDA safe minimum for a whole cut is 145 °F, 62.8 °C. So no: medium rare is below the minimum, by 5.6 °C, and this page says so plainly rather than pretending otherwise. The risk is real and it is small for an intact muscle from a reputable supplier, because the contamination on whole meat is on the outside and searing deals with the outside. It is not small for anything ground, tenderised, needled or brined, where the outside has been moved to the inside. That is why there is no safe rare burger and why this page refuses to offer a doneness scale for ground meat.

How much does meat keep cooking after it comes out of the oven?

Between about 2 °C and 15 °C, and the honest answer is that the published guidance and the physics disagree by roughly a factor of two. Certified Angus Beef tell you to pull a roast 5 to 10 °F below target and a steak 5 °F below. A conduction model — and the one published measurement of a thick steak — says more like 10 to 15 °F for a thick piece. The page gives you both, because the spread is the truth. What decides it is the shape, the surface temperature and how you rest it, and not, as most charts claim, the weight.

Does a bigger roast have more carryover?

Less than everybody says. Thickness enters twice and the two effects nearly cancel: a thicker piece stores more heat in its outer layers, and a thicker piece also loses a smaller fraction of it from its surface — but it also spends longer in the oven, so by the time it comes out the surface has reached the 100 °C boiling plateau and stopped getting hotter. In the finite-difference solve, changing the thickness by a factor of nine changes the delivered fraction by about a quarter. Shape does much more: a tied ball carries nearly twice the carryover of a flat steak at the same surface temperature.

Can I pull the meat early and let carryover take it to the safe minimum?

No, and this is the one thing on the page that is not a matter of taste. FSIS’s wording is that the food must reach the listed temperature as measured with a food thermometer before removing it from the heat source. Carryover cooking is a tool for landing on a doneness target that sits above the minimum without overshooting it. It is not a way of reaching the minimum. The page computes both numbers and labels the second one “lowest pull temperature that still reaches the USDA minimum on the heat” so you can see the difference.

Why does the three-minute rest appear in the 145 °F rule?

Because the rest is part of the lethality. USDA lowered the whole-cut figure to 145 °F with a three-minute rest in 2011; the pathogen reduction is a time-at-temperature product, and holding at 145 °F for three minutes delivers the same reduction as a higher instantaneous temperature. It is the only FSIS category with a rest written into the minimum, and the rest is not optional in the same way that the temperature is not optional.

Why does the oven temperature not change the answer?

Because above roughly 110 °C it genuinely does not, and almost no carryover chart admits it. What drives carryover is how much hotter the outside is than the middle when the meat leaves the heat, and while the surface is wet it cannot go above 100 °C — the water boiling there holds it. A 230 °C oven gets you to the same surface temperature as a 160 °C oven, just sooner. What does break the ceiling is a dry crust: a hard sear or a grill can put the surface at 120 to 160 °C, so put that number in the surface box and the estimate rises accordingly.

Related calculators

References

U.S. Department of Agriculture, Food Safety and Inspection Service. Safe Minimum Internal Temperature Chart and Doneness Versus Safety, fsis.usda.gov. A work of the United States Government and in the public domain. FSIS’s own wording is that the product must reach the listed temperature as measured with a food thermometer before removing food from the heat source — which is why this page will not treat carryover cooking as a way of reaching a minimum.
U.S. Food and Drug Administration. Safe Minimum Internal Temperatures as measured with a food thermometer, fda.gov/media/107000/download. The same chart, reproduced by FDA, used here as the cross-check on every figure: 145 °F with a 3-minute rest for beef, pork, veal and lamb chops, roasts and steaks; 160 °F ground meat; 165 °F poultry including stuffing; 140 °F for fully cooked ham from a USDA-inspected plant; 160 °F egg dishes; 145 °F fin fish; 165 °F leftovers and casseroles.
Certified Angus Beef. Degree of Doneness, certifiedangusbeef.com. The source of the rare-to-well-done scale used here — rare 125 °F/52 °C, medium rare 135/57, medium 145/63, medium well 150/66, and no well-done row at all — and of the pull-temperature rule this page defaults to: “remove from heat when the thermometer registers 5 to 10 °F lower than desired doneness” for roasts, 5 °F for steaks and burgers.
Meathead Goldwyn, Carryover Cooking: A Problem or Much Ado About Nothing?, amazingribs.com. The only published figure this batch could find that came from measuring rather than from a rule: a 1½ inch steak pulled at 130 °F reaching 145 °F after a ten-minute rest, and the observation that carryover grows with searing and with thickness.
The carryover model is this page’s own. The shape factor n/(n+2) follows from the quasi-steady conduction profile; the coefficient c was bracketed with an explicit finite-difference solve of the heat equation in slab, cylindrical and spherical coordinates (thermal conductivity 0.45 W/m·K, density 1050 kg/m³, specific heat 3400 J/kg·K, giving a diffusivity of 1.26 × 10⁻⁷ m²/s), with the surface held at 100 °C while wet and a convective boundary during the rest. It is an estimate and the page says so wherever it appears.
U.S. Department of Agriculture news release, 24 May 2011: “Cooking raw pork, steaks, roasts, and chops to 145 °F with the addition of a three-minute rest time will result in a product that is both microbiologically safe and at its best quality”, with the explanation that during the rest “the temperature stays the same or continues to rise, killing pathogens”. This is the revision that brought pork down from 160 °F to the 145 °F already applied to beef, veal and lamb, and that wrote the rest into the minimum. Ground meats and poultry were explicitly excluded and stayed at 160 °F and 165 °F.