Which wireless meat thermometer should you buy? Picks by how and where you cook

Three probes, three single-unit accuracy checks, and the results sit inside about a third of a degree of each other. AmazingRibs read a MEATER at 134.7, 159.9 and 205°F against 135, 160 and 205 (E·05), an RFX at 134.9, 160.2 and 204.9 (E·19), and a Combustion within 0.3°F from 100 to 210°F (E·14). Anyone hoping this guide would name the most accurate probe is out of luck. The record cannot.

It can say other things, and they matter more for the purchase. The three take different routes from the meat to your phone. One needs a spare phone to reach WiFi, one needs a separate accessory, and one will not work at all without its own base unit (E·03, E·09, E·16). Only one independent range test put them in the same closed grill and moved the receiver away, and it split them (E·23). So the picks below follow the cooker, the distance and what you want the app to tell you. Our MEATER Pro review (76/100, medium confidence) and two comparisons, RFX versus MEATER Pro and MEATER versus Combustion, hold the product-level detail.

Which MEATER, which Combustion, which RFX

Model names cause more mistakes in this category than any spec. Meater's support site says the MEATER Pro is the MEATER 2 Plus renamed, with the name as the only difference (E·01). The Pro is the one-probe model. The Pro Duo (two probes) and Pro XL (four) add, by Meater's account, built-in WiFi, but we found no independent test of either, so nothing here is claimed for them (E·02). The older MEATER Plus is a different, lower-heat product, and one owner thread we read compares a MEATER whose ambient limit matches that older probe rather than the Pro (E·25).

Combustion sells a second-generation Predictive Thermometer today (E·08). The independent accuracy check we found, and the only independent timing of its predictor, come from a review that mentions the newer release as a separate model, so we read them as prior-generation results (E·14, E·15). That review also says the second generation has WiFi built in, while Combustion's own spec table says WiFi needs an accessory; we follow the table (E·09, E·14). With the RFX, the probe and the Gateway are separate products, and the probe does not connect without the Gateway (E·16).

What the record separates, and what it doesn't

A tie on the first row is a finding. The same table shows where the gaps are, and all of them are in how the signal travels and what you must buy to receive it.

Range: the test that separated them

The Barbecue Lab ran nine probes through a ladder of closed-grill tests in a Weber Searwood, with the receivers moved farther away each time (E·23). Through the closed lid at 6 feet the MEATER Pro stayed connected. Through the lid and the whole house to the front porch, only two probes survived, and the RFX was one of them; the Combustion, whose kit included the Booster and Display, dropped. Inside a closed cast-iron Dutch oven at 6 feet, the RFX and the Combustion stayed up and the MEATER was not among the survivors (E·23). The lab's closing ranking put the RFX among the standouts and the Combustion a close runner up (E·24).

One owner thread points the same way. A user who ran all three reported that the RFX was "rock-solid" and that the Combustion had "connection issues early in the cook" that "settled down after about an hour" (E·25). That thread is an owner report, and its MEATER may not be the Pro, so it adds a direction, not a measurement (E·25).

This is one household, one grill and one Dutch oven, published in January 2025, and the Combustion kit's generation is not stated (E·22, E·23). The lab also moved base units and range extenders away from the cooker. For the MEATER that equipment is its charger, which Meater says to keep close to the cooker, so its dropouts may partly reflect that placement and not the probe alone (E·04, E·23). The lab's site carries merchant links and discount codes, but the codes we saw are for ChefsTemp and Typhur, not for these three products (E·22). The MEATER's maker numbers are no help in settling it: Meater's product page says 2,500 feet in open air, while its support guide gives 33 feet from probe to charger and 250 from charger to phone (E·04). The foot figures on the boxes are all maker numbers (E·04, E·09, E·17), which is why we lean on the test and not on the boxes.

Ambient: the wire is the difference

Two of the three put a temperature sensor in the handle. Meater says its sensor reads the air directly outside the meat and is not designed to confirm the cooker's set temperature (E·06). Combustion publishes a wide tolerance for its ambient sensor, ±2.7°F up to 212°F (E·08). ThermoWorks left the sensor out of the RFX probe and ships a wired air probe that clips to the grate (E·16, E·19). Our ambient explainer covers what the handle reading means.

The independent evidence is thin and points one way. The Barbecue Lab logged an 11-hour pork butt cook and found that for the first three hours no wireless ambient sensor came close to the grate temperature, sometimes sitting 50°F off, and that it took seven hours before any of them beat the RFX's wired probe (E·22). A smaller review, by a site that received the RFX free from ThermoWorks, checked a MEATER and an RFX with an infrared gun on a rotisserie turkey and found the MEATER about 100°F low (E·26). AmazingRibs read one MEATER handle sensor 7°F low at the ice point and 11°F high at the boil (E·05). Neither gives a figure for the Combustion, so its handle sensor sits in the same class without a number of its own (E·14, E·22, E·26).

The trade is a wire. A wireless ambient reading that stays accurate all cook did not appear in anything we read, and The Barbecue Lab's phrasing is blunter than ours: you can have a fully wireless setup or accurate ambient all cook, and apparently not both (E·22).

Predicted finish time is a feature, not a measurement

MEATER and Combustion both give an estimate, and neither maker's claim is a measurement. Meater says its estimate appears after the internal temperature has risen about 16°F, works best when the ambient temperature is steady, and does not account for wrapping or the stall (E·07). Combustion says its first countdown is usually within 10% of the true time and that on a low, slow cook it waits for the surface to pass the target (E·13).

Exactly one predictor has an independent timing. AmazingRibs ran the Combustion on a wet sock in a 325°F oven. The first prediction said 13 minutes and the cook ran seven minutes longer; a cook to a higher target temperature hit the stall and produced a much less accurate prediction, and AmazingRibs founder Meathead later reported predictions at higher grill and oven temperatures that were pretty good (E·15). That was probably the prior-generation probe (E·14). We found no independent timing of MEATER's estimator (E·07), and ThermoWorks' app showed no finish-time feature in the release notes and the review we read (E·20, E·21). If the estimate is your reason to buy, the evidence supports a feature, not a ranking.

Hardware you pay for, and software you depend on

The RFX is the only one of the three that cannot work as a lone probe. It needs its Gateway, which comes in starter kits and must be bought if you only buy a probe (E·16). The Combustion probe is Bluetooth-only; WiFi arrives through a Booster, a Display or a Giant Grill Gauge, and on 28 October the Booster's page showed a November ship date (E·09, E·10). The MEATER Pro reaches WiFi through a second phone or tablet that has to stay within Bluetooth range for the whole cook, or through the Duo or XL (E·02, E·03). Each route means buying, charging and placing something, and the MEATER route costs nothing only if you already own the spare device.

Three separate questions sit under software. Current function is fine on all three: each app's store listing showed recent updates (E·21, E·28, E·29). The dependency architecture differs. The RFX needs the ThermoWorks app and an account, free for up to 10 devices, and has no physical alarm, so alerts are app notifications; a Bluetooth mode removes the WiFi requirement (E·18). MEATER's WiFi route needs a MEATER Cloud account (E·03). Combustion's app is free, its libraries are open source under an MIT license, and its Display shows a countdown without a phone (E·11, E·12). Open code reduces lock-in; it does not guarantee a cloud service (E·12). Past app shutdowns elsewhere, like the one in our Joule guide, show the risk class exists and say nothing about these three.

Pick the ThermoWorks RFX if the signal has to get out of something

Choose this if the probe sits inside a closed steel smoker, a Dutch oven or a thick-walled cooker, or the receiver will be a floor or a house away. The evidence is the range ladder, the owner thread and the design: a 433 MHz radio to a Gateway that the maker rates for 560 feet obstructed, and a wired ambient probe (E·16, E·17, E·19, E·23, E·25). It is also the pick if you want trustworthy ambient numbers and can live with a wire (E·22).

The cases against it are specific. Wrapped in aluminum foil, one owner reported both probes dropping, and later retests under foil failed again; ThermoWorks' own forum account says foil can act as an RF barrier and recommends pushing the probe through the wrap (E·27). Nothing we read says the MEATER or the Combustion do better under foil. The Gateway is mandatory, WiFi is 2.4 GHz only, and the app and account are required (E·16, E·17, E·18). The finish-time feature is missing (E·20, E·21), and the probe warranty is one year (E·16).

Pick the Combustion if you want the estimate and the most sensors

Choose this if you want the probe that finds the coolest point for you, with eight sensors and a feature that finds the lowest internal temperature even if you miss the center (E·08). Add the Display if you want a countdown on the counter without a phone, or if you run up to four probes (E·11). It passed the Dutch oven test and trailed only on the whole-house run (E·23, E·24).

Against it, the accuracy record is a generation old and the one independent timing was seven minutes off early (E·14, E·15). WiFi needs extra hardware, and its range, which the maker lists as 10 to 100 feet obstructed, failed across the house in the one test (E·09, E·23). The owner thread reports early connection trouble, at low confidence (E·25).

Pick the MEATER Pro if the cook happens near the phone

Choose this for an oven or an open grill within roughly a phone's reach, one cut at a time, and you want nothing else to buy. It needs no hub or second device for the basic setup, though its charger is part of the signal path and has to sit close to the cooker (E·02, E·04). The independent food reading was as close as the others, and one reviewer called its app the friendlier one (E·05, E·26). Our review of the MEATER Pro holds the rest.

It is the one that dropped in the range test, at a Dutch oven and across the house (E·23). Its handle sensor read 7°F low and 11°F high in one check and about 100°F low on a turkey (E·05, E·26). Its estimate does not account for a stall or a wrap (E·07), its maker range numbers describe different legs of the link (E·04), and WiFi means a second device or a different model (E·03).

Profiles where none of these wins

A cook who wants a fully wireless setup and an accurate ambient reading all cook has no match here, because the only ambient reading that held up in the one test came from the probe with a wire (E·19, E·22). A pitmaster who wraps in foil and wants one wireless probe under the wrap has no documented answer: the RFX has a documented failure, and the other two have no evidence either way (E·27). And anyone buying for a finish time they can plan dinner around has a feature to try and no verified predictor to trust (E·15).

For a quick check of doneness, none of these is the right tool; our instant-read guide covers that.

What would change these picks

A second independent range test with distances and probe models stated would either confirm the ladder or break it, and it would settle whether the Combustion's house failure repeats on today's hardware (E·23). An independent accuracy check of the second-generation Combustion would close the gap in its record (E·14). A ThermoWorks finish-time feature, or an independent timing of MEATER's estimator, would reopen the prediction row (E·15, E·20). And a foil test of the other two would show whether the RFX's problem is unique (E·27).

The range ladder is one household's test, every accuracy figure is one unit, the Pro Duo and Pro XL have maker claims only, and the owner reports come from single cooks (E·05, E·19, E·23, E·25). Prices and ship dates move, so check them where you shop.

The bottom line

Buy the RFX if the probe will sit in steel or the receiver will be far away, and accept the Gateway, the account and the foil limit. Buy the Combustion if the estimate and the sensor count matter more than reach, and add a Booster or Display for WiFi. Buy the MEATER Pro if you cook within phone range and want the simplest start. Don't buy any of them for a finish time you plan dinner around, or for a fully wireless ambient reading.