2024 test14 productsHand Tools
Which Flare Nut Wrench Brand Wins?
We compared 14 flare nut wrench options head to head. Snap-on came out on top. See the measured results, the runner-up, the budget pick, and a link to the full test video.
Some figures on this page were transcribed from the test video and have not been independently re-verified. Treat the numbers as a close guide and watch the full video for the exact readings.
Winner
Snap-on
Price shown in test: $285 for 5 wrenches (most expensive brand tested)
Check price on Amazon
Affiliate link. We may earn a commission at no extra cost to you.
Runner-up
Proto
Price shown in test: not stated verbatim; the price-introduction sentence for this brand appears to have been dropped by captions (the transcript jumps directly from another brand's failure-load number into "in USA and it's 122 G for the Proto" with no dollar figure). The narrator's closing statement that "for less than half the price the Proto performed better where it counts" implies a price under half of Snap-on's $285, but no exact figure is ever spoken
Check price on Amazon
Affiliate link. We may earn a commission at no extra cost to you.
Budget pick
Williams
Price shown in test: $37
Check price on Amazon
Affiliate link. We may earn a commission at no extra cost to you.
The measured results
Every number below is read straight from the test. Scroll sideways to see all measurements. Products are listed in the order they finished.
| Product | Weight/head/slop (representative wrench) | 3/8 in fitting torque-to-failure test | 1/2 in coupling nut test, attempt 1 | 1/2 in coupling nut test, attempt 2 | 5/8 in coupling nut test | Overall average finish | 1/2 in coupling nut test | 5/8 in coupling nut test, attempt 1 | 5/8 in coupling nut test, attempt 2 |
|---|---|---|---|---|---|---|---|---|---|
| 1Snap-on Flank Drive$285 for 5 wrenches (most expensive brand tested) | 112 g, head 0.7335 in (most compact of all 14 brands), 0.465 in of slop (best/tightest fit of all 14 brands) | 271.50 in lb, 2nd place behind Proto | approximately 912 in lb (transcript literal digits read "92.8" in two separate mentions; corrected here because both mentions independently frame it as "only 3 in lb less than" Proto's 915.6 in lb break, and a dropped leading 1 explains "92.8" reading as 912.8) | over 700 in lb per the closing summary ("the Snap-on flank drive wrench outperformed all of the flare nut wrenches on the second attempt however the Williams and Snap-on flare nut wrenches performed almost as well at over 700 in lb"); transcript's raw digit string for this figure reads "79.6," likely missing a leading digit (roughly 796), kept flagged rather than corrected | 1,282 in lb then 1,241 in lb on a second attempt on the undamaged part of the same nut, only a 3.7 percent drop, no visible wrench damage but heavy nut damage; a later closing recap sentence for this size is too garbled to cleanly parse ("the Snap-on flare nut wrench finished in third at 1,128 in pound") and is not treated as reliable | 1.71, best/declared overall winner | not tested | not tested | not tested |
| 2Protonot stated verbatim; the price-introduction sentence for this brand appears to have been dropped by captions (the transcript jumps directly from another brand's failure-load number into "in USA and it's 122 G for the Proto" with no dollar figure). The narrator's closing statement that "for less than half the price the Proto performed better where it counts" implies a price under half of Snap-on's $285, but no exact figure is ever spoken | 122 g, head 0.785 in, 0.533 in of slop | 289.50 in lb, best of all 14 brands | approximately 915.6 in lb (transcript reads "9156," read here as a garbled decimal consistent with this test's other torque figures), wrench broke ("career ending injury"), the highest first-attempt figure recorded | not tested | approximately 1,350 in lb then 1,270 in lb on a second attempt (transcript reads "1,35" for the first figure, likely missing a trailing zero); narrator states only a 2.7 percent drop between attempts, which does not cleanly match the literal 1,350 to 1,270 figures (a 5.9 percent drop by that math), flagged as an unresolved inconsistency rather than corrected | not tested | not tested | not tested | not tested |
| 3Williams$37 | 92 g, head 0.838 in, slop given as "0.5 3 4 in" in the transcript (ambiguous between 0.534 in and 0.574 in; kept as flagged rather than guessed) | 246.50 in lb, called out in the recap as performing well, close to Kobalt | approximately 848.3 in lb (transcript reads "8483"), moved into the lead at that point | approximately 755.3 in lb, the best second-attempt figure recorded; corrected from the transcript's garbled "700 55.3" because the narrator's stated 11 percent performance drop from 848.3 in lb computes to about 755, matching | approximately 1,130 in lb then 832.4 in lb on a second attempt (transcript reads "1,13" for the first figure, likely missing a trailing zero); narrator frames this as continuing to outperform most of the competition | not tested | not tested | not tested | not tested |
| 4Crescent$16 | 116 g (heaviest brand recorded up to that point in the video), head 0.774 in (most compact recorded up to that point), 0.577 in of slop (least slop recorded up to that point, though a smaller wrench size than Efficere/Duratech) | 196.7 in lb, moved into the lead at that point | approximately 881.9 in lb (transcript reads "8819"), only barely outperforming a competing brand | approximately 731.9 in lb, derived from the narrator's stated "lost about 150 in lb of performance on the second attempt" rather than a verbatim second figure; not itself a spoken number, flagged as an approximation | recorded only in a heavily garbled closing recap segment for this size; not extracted with confidence | not tested | not tested | not tested | not tested |
| 5Efficere$10 (least expensive brand tested) | 102 g, head 0.855 in, slop described as "almost 3/4 of an inch," a very loose fit | not recoverable; the transcript sentence introducing this brand's own slip/failure point on the 3/8 in test appears to have been garbled and dropped entirely ("any fser slipped at is made in China and the durch weighs 94 G" runs two sentences together with no clean number for Efficere) | 710 in lb | 506 in lb, a loss of over 200 in lb of strength, consistent with the stated numbers | an initial reading of about 8.5 in lb was recorded, then discarded and retested on a brand new coupling nut at 43.2 in lb; both figures read as very poor performance for this brand | not tested | not tested | not tested | not tested |
| 6Duratechnot stated verbatim; the price-introduction sentence for this brand appears to have been dropped by captions between Efficere's section and the statement "is made in China and the durch weighs 94 G" | 94 g, head 0.874 in, 0.598 in of slop | 194.7 in lb, outperformed Efficere | approximately 668.3 in lb (transcript reads "6683") | approximately 622 in lb (transcript reads "622.org," the ".org" is a caption artifact), jaw damage measured at 8.02 mm | approximately 798 in lb (transcript reads "798.org") | not tested | not tested | not tested | not tested |
| 7Performance Tool$13 | approximately 100 g, head 0.794 in (about 10 percent smaller than Duratech's), slop described as more than the dial indicator could measure, well over an inch, the worst build-quality result of the video's early brands | gave up before the dial indicator's 60 in lb minimum threshold was even reached, described as "a very poorly designed wrench with this much taper" | 836 in lb, outperformed Efficere and Duratech | approximately 688.8 in lb; transcript reads "6888," read here as a likely dropped decimal since Crescent's own second attempt is separately described as "just over 688 in lb" while performing a similar amount worse than its first attempt as Performance Tool, though this is not a fully certain correction | approximately 757.8 in lb then 686 in lb on a second attempt (transcript reads "7578" and "686.com"); the stated 9 percent drop from 757.8 computes to about 690, closely matching 686 | not tested | not tested | not tested | not tested |
| 8Kobalt$20 | 100 g, head 0.879 in (a notably big head), 0.61 in of slop, described as having "quite a bit of taper" | approximately 247.3 in lb (transcript reads the badly garbled string "247.374410"), the first flare nut wrench in the video to outperform the open-end Snap-on baseline | not tested | not tested | recorded as the garbled string "868.42mhz" for a first figure and "32.9" for a second; narrator states this brand "performed about as well as the Crescent" (roughly 700 to 880 in lb range), so the literal "32.9" figure is very likely missing a leading digit, flagged rather than corrected | not tested | recorded as approximately 691.5843 (transcript caption split the decimal across a line break: "691.584" then "3 mm"; joined as one continuous reading). The trailing "mm" unit is likely a captioning artifact: no other coupling-nut reading in this segment reports a jaw-growth figure anywhere near this magnitude (compare 8.29mm, 9.01mm elsewhere), while 691.58 sits squarely in the same range as this brand's own torque-to-failure readings (247.3, and the 5/8in test's ~700-880 range); most likely this is an in-lb torque figure with a mislabeled unit, but not reassigned here since the unit itself was not part of the split-decimal glitch. | not tested | not tested |
| 9Craftsmannot stated verbatim; the price-introduction sentence for this brand appears to have been dropped by captions between the Kobalt section and "the Craftsman head is a little bit large at 0.8335 in" | head 0.8335 in (a little large), 0.585 in of slop, similar to Kobalt | gave up under 95 in lb, described as "way too much taper," only making contact with around 60 percent of the fitting; the weakest performer on this test | approximately 477.2 in lb (transcript reads "4772") | 386.3 in lb, described as a "season ending injury," jaw contact area measured at 8.29 mm but the jaws described as too weak to work effectively | recorded as 71.1 in lb for a first attempt and approximately 699 in lb for a second attempt (transcript reads "699.00 B"), which the narrator describes as "performed about the same on a second attempt"; this framing is inconsistent with the literal first-attempt figure of 71.1 being so much lower, so the 71.1 reading is flagged as a likely caption error rather than corrected to a specific number | not tested | not tested | not tested | not tested |
| 10Quinn$45 for 5 wrenches ($9 each), sold at Harbor Freight | 94 g, head 0.8375 in (almost the same as Williams), 0.602 in of slop (more than Williams) | transcript reads "29.3 in lb," which is roughly 10 times lower than every neighboring brand's result in the 190 to 290 in lb range on this same test; flagged as a likely dropped leading digit (plausibly closer to 200 to 290) rather than corrected to a specific value | approximately 559.1 in lb (transcript reads "559.115"), described as "a bench warmer performance" | not given as a verbatim figure; the narrator states only a 28.3 percent drop from the first attempt, which would put the second attempt at roughly 401 in lb, but this is a derived approximation, not a spoken number | approximately 661.8 in lb then 647 in lb on a second attempt (transcript reads "6618" for the first figure) | not tested | not tested | not tested | not tested |
| 11GearWrench$45 for six wrenches, stated in the transcript as "$8 each"; the arithmetic does not divide evenly (45 divided by 6 is $7.50, not $8), flagged as a possible garbled price rather than corrected | 110 g (heavier than average), head 0.7865 in, 0.766 in of slop, described as "way too much slop" for its otherwise low-taper design | 196 in lb, described as "still pretty decent" despite the excess slop | 698.4 in lb, described as performing about the same as Efficere | approximately 652.1 in lb (transcript reads "6521"); the narrator's stated 6.6 percent drop from 698.4 computes to about 652.5, closely matching | recorded as the garbled string "856.011 15.6 in lbs," roughly 856 in lb based on the leading digits, not extracted with full confidence | not tested | not tested | not tested | not tested |
| 12Milwaukee$46 for four wrenches (about $12 per wrench) | 110 g, head 0.778 in (more compact than average), 0.542 in of slop (a pretty snug fit) | transcript reads "1991.5 in lb," implausibly high compared to every other brand's 95 to 290 in lb range on this same test; flagged as a garbled/misplaced-decimal figure rather than corrected to a specific value | transcript reads "74.2 in lb," implausibly low compared to the roughly 560 to 915 in lb range recorded for every other brand on this test; flagged rather than corrected | transcript reads "585.3 in lb" (from "5853") with a stated 22.4 percent drop from the first attempt; the two figures do not reconcile cleanly with either possible reading of the first-attempt number, so both are kept flagged rather than resolved | recorded as the garbled string "776.012," roughly 776 in lb based on the leading digits, not extracted with full confidence | not tested | not tested | not tested | not tested |
| 13Matco Tools$72 for four wrenches ($18 each) | 118 g (heaviest recorded up to that point), head 0.768 in (most compact recorded up to that point), 0.516 in of slop (tightest fit recorded up to that point) | approximately 188 in lb (transcript reads the garbled "188.jpg"), described elsewhere as an "about average" first-round result relative to the field | approximately 466.4 in lb (transcript reads "4664"), described as "a pretty dismal" result | 42.8 in lb, described as throwing in the towel very early again, a dramatic decline from the first attempt | not tested | not tested | not tested | approximately 698.4 in lb (transcript reads "6984"); cross-checked against the narrator's statement that this is "100 in lb less than" Duratech's own 5/8 in figure of about 798, which matches closely | recorded as "65.2 in lb" against a stated "almost 12 percent" drop from 698.4 (which would compute to about 615); the literal figure and the computed figure do not agree, flagged as a likely misplaced decimal rather than corrected |
| 14Mac Tools$250 for five wrenches ($50 each) | 138 g (heaviest of all 14 brands), head 0.749 in (most compact of all 14 brands), 0.639 in of slop (a pretty sloppy fit despite the compact head) | 145 in lb, described as "really struggled on this test," with "way too much taper and way too much slop to compete with the best in the showdown" | 722.9 in lb, described as "isn't too impressive considering the price of the tool" | 632 in lb, a 12.6 percent drop from 722.9, which matches the stated percentage closely | not tested | not tested | not tested | approximately 1,430 in lb (transcript reads "1,43"), described as performing well above average | approximately 928.7 in lb (transcript reads "9287") |
How it was tested
- brand weight and head size of the most common representative wrench in each set (grams and inches)
- shaft/jaw slop or backlash measurement with a dial indicator (inches)
- 3/8 in brake line fitting torque-to-failure test using a calibrated Proto torque wrench (in lb), including an unranked GearWrench regular open-end wrench as a baseline control
- 1/2 in soft coupling nut torque test, two attempts per brand on rotated/fresh contact points (in lb and mm of jaw damage), including an unranked GearWrench regular open-end wrench as a baseline control
- 5/8 in soft coupling nut torque test, one or two attempts per brand depending on nut damage (in lb and mm of jaw gap growth), with no open-end wrench baseline available at this size
- overall average finish/ranking across the flare nut wrench tests
“The Snap-on flare nut wrench set is extremely expensive but it came out on top with an average finish of 1.71.”
Data notes and caveats
14 flare nut wrench brands (Snap-on, Proto, Williams, Mac Tools, Crescent, Kobalt, GearWrench, Duratech, Milwaukee, Performance Tool, Craftsman, Matco Tools, Quinn, Efficere) tested at three fitting sizes (3/8 in, 1/2 in, 5/8 in), plus an unranked GearWrench regular open-end combination wrench used throughout as a control to test the video's second research question: whether a standard open-end wrench can match a flare nut wrench. The narrator's explicit conclusion on that question: flare nut wrenches keep their advantage on small brake-line-sized fittings (3/8 in) but lose it as fastener size increases, with a good open-end wrench being very hard to beat at 5/8 in and larger. This transcript is exceptionally garbled with dropped decimal points, missing leading/trailing digits, and stray caption artifacts (".org", ".com", "mhz", "/ MIM") scattered across dozens of individual torque figures, especially in the 1/2 in and 5/8 in sub-tests; every questionable figure is flagged in its brand's results rather than silently corrected, and confidence is set to low accordingly, even though brand identity resolution (Efficere, Duratech, Kobalt, Matco Tools vs Mac Tools) is fairly solid. No meta chapters exist for this video. The narrator's stated overall winner by average finish is Snap-on (1.71), but the narrator's own explicitly stated personal favorite is Proto, and both Williams (best value) and Crescent (best money-saver) receive separate distinct value callouts near the end.