Anova Precision Cooker 3.0 vs Breville Joule Turbo: what the fastest circulator asks in return

One of the best-reviewed kitchen gadgets of the 2010s stopped working for some of its owners this spring, and no part of the hardware failed. The company retired the app the device shipped with, announcing the date in a Facebook post. The machine has no buttons, no dial, no display. The app was never a remote control; it was the only control. Owners who moved to the successor app kept cooking. At least one reported that the required migration and firmware update ended in a blinking red light and a dead unit. That event hangs over this comparison, because the fastest immersion circulator in the measured data is built on the same plan.

The Breville Joule Turbo preheated a 9.5-quart bath four minutes faster than the Anova Precision Cooker 3.0 on the same bench, 32 minutes against 36 (E·04), and it runs entirely from a phone app (E·08). The Anova is the category default, and it keeps every cooking function on its own touchscreen (E·01). We didn't heat any water ourselves. We read both companies' pages, pulled TechGearLab's instrumented measurements, traced the app record on both sides, and went through what happened to the machine that lost its software in March. This piece prices the measured speed against what the faster machine asks you to accept.

The short table

Values from the manufacturers' pages and TechGearLab's instrumented tests (E·01, E·02, E·03, E·04, E·05, E·06, E·07). The heat-up row is the Turbo's win. The rest of the table is why it isn't the default.

Four minutes, and a claim of more

Both machines carry 1100 W heaters, and the same lab timed both on the same 9.5 quarts of 50°F tap water (E·04). The Joule Turbo reached 147°F in 32 minutes, the best time of the seven units tested (E·04, E·05). The Anova followed at 36, second place (E·04, E·07).

Measured data

Four minutes on a half-hour preheat is real, and it is also the entire measured gap. Start from hot tap water and every bar in that chart shrinks together. If your weeknights are tight, four minutes is four minutes. It is not a different life.

Turbo mode is Breville's argument that the gap is much bigger. The company markets 'Perfect results in half the time' on select recipes, crediting algorithms it does not explain further (E·03), and the app's Autopilot copy claims proteins cook up to twice as fast as traditional sous vide (E·11). We found no instrumented timing of Turbo mode anywhere in the lab data, and TechGearLab documented its boundary instead: Turbo works only inside Breville's own recipes, and you cannot apply it to a cook you design yourself (E·09). Cook the app's standard steaks and chicken, and the half-time claim covers you, on Breville's word. Cook from your own time-and-temperature tables, and the Turbo's advantage falls back to the measured four minutes (E·04, E·09). Nobody outside Breville has clocked Turbo mode against a stopwatch (E·03, E·09, E·11), so how much time it actually saves on food you cook yourself stays the company's claim to make.

Accuracy, from the same bench

Speed was the Turbo's row. This one belongs to the Anova, and neither number may be waved away, because both came out of the same lab under the same protocol. TechGearLab logged the Anova's bath at 0.06°F from setpoint on average, stable across a full hour (E·07). The Joule Turbo ran 0.45°F cooler than setpoint on average, and did it consistently (E·06).

Half a degree cool is not a scandal. A consistent offset can be canceled by setting the target half a degree high, once you know your unit has one (E·06). But an error you correct is still an error you are managing on the machine's behalf, and the setpoints where sous vide earns its keep, the edge of a 129°F steak or a 145°F custard, are decided by amounts about that size. A bath measured at six hundredths of a degree needs no bookkeeping (E·07).

The caveat cuts both ways: each figure describes one retail unit, and this lab has caught two revisions of another brand's circulator diverging under the same protocol (E·06). The distance between 0.06 and 0.45 is wide enough to survive some unit scatter. It is not a promise about your serial number.

The lab's own totals put the Anova first of seven at 94 and the Turbo third at 84, with a Top Pick award attached (E·05, E·07). Both machines left that bench liked.

What still works when the phone doesn't

Three separate things are true about the Joule Turbo and its app, and they stay separate here.

The first is about today, and today the picture is fine. The Breville+ app is free on its store listing, with no subscription tier or in-app purchase shown (E·11), but TechGearLab's review of the Turbo, dated December 2024, says full access to specific chef recipes requires a subscription (E·19). We could not reconcile the two, so whether any part of Breville+ is paid is unresolved. The lab that docked the Turbo for app dependence still handed it an award (E·05), and nothing in the current record shows Turbo owners unable to cook.

The second is architecture, and it does not depend on anyone's intentions. The Turbo has no on-device controls of any kind (E·08). It pairs with the Breville+ app and with nothing else; a ChefSteps support note from September 2023 confirms it will not pair with the older Joule app (E·10). Voice control through Google or Alexa exists (E·03), a second interface but not an independent one, since it rides the same cloud. Every cook, for the life of the machine, runs through a working app, a signed-in account and a phone in the house.

The third is history, and it belongs to a different machine. In March 2026 the original Joule's app was shut down, with owners pointed to Breville+ (E·14). That machine was also app-only (E·15), so the migration carried absolute stakes, and the record holds at least one unit that did not survive it: working before the required firmware update, blinking red after, with support offering a discount on a replacement rather than a fix (E·16). ChefSteps documentation for the original also lists a quieter failure, a unit that stops circulating when its WiFi connection drops repeatedly, with Bluetooth-only control as the workaround (E·17). Owners navigating that migration today can find the documented failure pattern and the fallback options laid out in our guide to the original Joule's app shutdown.

That history proves less than it might seem to. It does not show the Turbo will lose its app; Breville migrated original-Joule owners onto a maintained replacement, and the Turbo's app is that replacement (E·10, E·14). What the record establishes is that circulator apps get retired, that forced migrations can hurt, and that the damage depends on the architecture underneath. Nobody, including Breville, has put a date on how long the Turbo's own app gets that same maintenance; the policy is silence in both directions, not a promise (E·10, E·14).

Anova belongs in this section too, on the other side of the ledger but not off it. Anova put parts of its app behind a fee for new accounts in August 2024, $1.99 a month or $9.99 a year, and at the time said older accounts would be grandfathered as an intent, not a contract; Anova's current apps page does not repeat that promise (E·12, E·13). The difference is what the fee can reach. Anova states its devices work without any subscription, and every model except the Mini keeps manual controls (E·13). Its plan table lists remote temperature and time control, progress monitoring and remote start and stop as free, and puts the recipe library, custom recipes, sharing, cook history and push notifications in Premium (E·20). When Anova monetized its platform, an owner with buttons kept the machine and its remote control, and only the recipe and history layer moved behind the fee. If Breville ever does the same, or worse, a Turbo owner has no buttons to fall back on (E·08).

Long cooks

Past hour one, no independent data exists for either machine, so the evidence is owner reports and design features. On the Anova side the record says lid, firmly: one reported 24-hour cook at 180°F evaporated down to the low-water sensor around hour 14 and shut off, exactly as designed (E·18, E·01). The 3.0 brings that 65 mm shutoff, an IPX7-sealed head and a 2-year warranty to the job (E·01); its longer reliability story, including one documented out-of-warranty death, is walked through in our Anova 3.0 review.

The Turbo's side of this section is mostly blank. We found no published long-cook measurements for it, no known-issues page, and no owner-failure cluster; the machine is younger and its record thinner, and long-term reliability on either side of this comparison is thin for the same reason: neither circulator has been on the market long enough to generate one. The one documented connectivity failure in the family, circulation stopping when WiFi drops repeatedly, is written about the original Joule, and we could not determine whether the Turbo inherits it, or what it does mid-cook if its own connection drops (E·17). For a 48-hour cook, that unknown sits differently on a machine you can only reach through the network (E·08).

What the money buys

List against list, the Anova sat below the Turbo on 28 October 2026 (E·02, E·03). Earlier in October Anova's own store also showed a sale price under list; that sale was gone on 28 October, so we no longer treat any sale figure as a working price, and Anova has run sales before, so check the current price (E·02). The Turbo was out of stock on Breville's US page that day, so its price is moot until it returns (E·03). We did not check retailer prices or stock.

Containers settle the rest. The Anova ships with its adjustable clamp, measured at a 1.4-inch gap and praised in testing, which grabs stockpots, tubs and coolers (E·07, E·01). The Turbo ships with a clip the lab called narrow and limiting, plus a magnetic foot that is genuinely clever on magnetic steel and induction-ready pots (E·05, E·03). Cook in a polycarbonate tub or a cooler and the Turbo needs the Big Clamp, sold separately (E·03). For a tub cook the Turbo's total therefore sits further above the Anova's than the list gap alone.

What would change this

• On-device controls on a future Joule, or any documented local-control path for the Turbo. The dependency half of this comparison would collapse overnight (E·08, E·10).
• A dated, written Breville commitment on how long the Turbo's app will be supported. Policy prose is weak evidence, but it would beat the current silence (E·14).
• An instrumented test of Turbo-mode cook times. Right now half-the-time is the manufacturer's claim, bounded to its own recipe catalog (E·03, E·09, E·11).
• Multi-unit accuracy data on either machine. The 0.06°F and 0.45°F figures each describe one unit (E·06, E·07).
• Anova gating any device-side function behind its subscription. Its written policy says it won't (E·13), and the Anova case here leans on that staying true.
• The Turbo back in stock at a price at or below the Anova's, with the Big Clamp in the box; the value gap would close to the four minutes and the offset (E·02, E·03, E·04).
• A clear statement of whether any Breville+ feature is paid; the lab and the store listing currently disagree (E·11, E·19).

Weigh those gaps against what's actually measured, not against the marketing on either box. This comparison rests on single-unit lab figures and a shutdown record that belongs to a different machine than the one being sold today, and neither circulator has been around long enough for a real reliability history to form. If neither machine fits, the cheapest-competent conversation and the very-large-vessel conversation both live in the circulator guide.

Update, 28 October 2026: an earlier version called the Anova's earlier sale its routine price and sent readers to the Joule Turbo as a live pick. Anova's store showed no sale on 28 October, Breville's US page showed the Turbo out of stock, and we found that the lab reports a paid tier for some Breville+ recipes that the store listing does not show. Dollar figures now sit in the evidence ledger with their dates.