Which Oil Stop Leak Additive Brand Wins?
We compared 14 oil stop leak additive options head to head. AT-205 came out on top. See the measured results, the runner-up, the budget pick, and a link to the full test video.
AT-205
Price shown in test: $20
Affiliate link. We may earn a commission at no extra cost to you.
Seal Lube
Price shown in test: $35, tied with Hapco as the most expensive product tested
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 | Mixing Behavior | Lubricity Wear Scar | Evaporative Loss Hardened Test | Hardened O Ring Thickness Restoration | Hardened O Ring Hardness Restoration | Gasket Puncture Force | Cold Flow Test | New O Ring Swell Test |
|---|---|---|---|---|---|---|---|---|
| 1AT-205 Re-Seal$20 | very cloudy when mixed with oil, separated and settled at the bottom after 5 minutes, remixed before testing | 8.42 mm versus the 7.05 mm straight motor oil control; energy use meter over 570 W, test pin very hot and blowing off steam | experienced a lot of evaporative loss, 2.21 g, among the higher losses of the lineup | 0.2055 in (from a hardened baseline of 0.194 in, versus an original healthy 0.206 in); came out on top of this category | lowest (softest) hardness score of any product at 65.5, versus a hardened baseline of 89.5 and an original healthy value of 75; described as actually softer than new; came out on top of this category | 430 g, the best (softest) result of any product, versus a heat-hardened control of 798 g; came out on top of this category | separated from the oil and went down the cold-flow slide way too quickly, a poor result despite winning the other categories | caused a healthy, unheated O-ring to swell to 0.215 in (from a 0.207 in control) and its hardness to drop from 75 to 70, the largest swell/softening effect on a healthy seal of any product tested |
| 2Seal Lube$35, tied with Hapco as the most expensive product tested | pretty cloudy when mixed, separated from the motor oil after a couple of minutes, remixed before testing | no result found in the transcript; Seal Lube does not appear in the initial lubricity wear-scar test at all, a probable gap (see videoNotes) | 1.32 g loss, described as a bigger loss than Hapco's | 0.2025 in, second-best result behind AT-205 | 73, second-best (second-softest) result behind AT-205's 65.5 | 446 g, second-best result behind AT-205's 430 g | separated from the oil and went down the cold-flow slide quickly, a poor result | caused a healthy, unheated O-ring to swell to 0.215 in (tied with AT-205 for the largest swell) and its hardness to drop to 72.5 |
| 3Lubegard Seal Fix$16, tied with CD-2 | became very cloudy, separated from the motor oil and settled at the bottom after 5 minutes, restirred before testing | 8.07 mm, the most damage of any product up to that point in testing | 1.83 g loss, the most of any product up to that point in testing | 0.2015 in, third-best result behind AT-205 and Seal Lube; narrator: 'the Lubegard actually works, the O-ring did recover some of its original size' | fully restored to the original healthy value of 75, fourth-best (per the final recap listing AT-205, Seal Lube, Justice Brothers, then Lubegard) | 528 g, described as the best result up to that point in testing at the time | separated from the oil and quickly went down the cold-flow slide, a poor result | caused a healthy, unheated O-ring to swell to 0.2125 in and its hardness to drop to 71.5 |
| 4Bardahl No Smoke Stop Leak$18, tied with Lucas | very high viscosity, reminded the narrator of Motor Honey, no separation from the motor oil | 6.52 mm, the best (lowest) result of any product tested, outperforming straight motor oil's 7.05 mm baseline | completely unfazed by the heat, no measurable loss, the best result of any product | 0.196 in, did not help much versus the 0.194 in hardened baseline | 88, only a slight softening from the hardened baseline of 89.5 | 724 g, a little better than CD-2's 760 g but not a large improvement over the 798 g control | had a huge impact on cold oil flow performance, a poor result | 0.209 in, hardness unchanged at 75 |
| 5Liqui Moly Motor Oil Saver$14 | no separation from the motor oil after 5 minutes | 6.85 mm, outperformed the straight motor oil control (7.05 mm), less friction than the control per the energy use meter | 0.63 g loss | 0.1965 in, similar to Bar's Leak, little improvement over the 0.194 in hardened baseline | 88, a small softening from the hardened baseline of 89.5 | 682 g, described as the best result yet at that point in testing | froze and moved down the cold-flow slide like an egg yolk, a poor result | 0.208 in, hardness unchanged at 75 |
| 6Justice Brothers Engine Stop Leak$22, tied with BlueDevil | became very cloudy but did not separate after 5 minutes | 9.33 mm, tied with CD-2 for the most damage of any product tested | boiled off 8.78 g, described as the most evaporative loss of all the brands tested; oil blend was observed actively boiling at around 200 F during the heat exposure | 0.199 in, better than average | 74, very close to the original healthy value of 75, third-best result behind AT-205 and Seal Lube | 542 g, moved into third place behind AT-205 and Lubegard at that point in testing | moved a little faster than the motor oil control, but viscosity seemed acceptable | 0.213 in, hardness softened to 71 |
| 7BlueDevil Oil Stop Leak$22, tied with Justice Brothers | became very cloudy, separated from the oil like Lubegard and AT-205, remixed before testing | 8.53 mm, quite a bit of damage; described as blowing off steam like Justice Brothers | 1.53 g loss | 0.199 in, same as Justice Brothers | 80, softened from the hardened baseline of 89.5 | 526 g, very close to Lubegard's result | separated from the motor oil, a poor result | 0.213 in, same as Justice Brothers, hardness 71.5 |
| 8Hapco Tri-Seal$35, tied with Seal Lube as the most expensive product tested | became very cloudy but did not separate after 5 minutes | no result found in the transcript; Hapco does not appear in the initial lubricity wear-scar test at all, a probable gap (see videoNotes) | 0.78 g loss | 0.197 in, described as some improvement | 81.5, softened from the hardened baseline of 89.5 | 492 g, described in the transcript as '200 g better than the control' though the 798 g heat-hardened control minus 492 g is actually a 306 g difference; the verbatim spoken comparison is preserved rather than corrected; third place behind AT-205 and Seal Lube in the final recap | flowed a little faster than the motor oil control | 0.213 in, hardness softened to 71 |
| 9CD-2$16, tied with Lubegard | no cloudiness when mixed, did not separate after 5 minutes | 9.33 mm, tied with Justice Brothers for the most damage of any product tested at that point ('by far the most yet'); test pin very hot to the touch | 0.05 g loss, a good (low) result | did not help the O-ring expand in size; no specific thickness figure given in the transcript for this brand | 85.5, only a small softening from the hardened baseline of 89.5 | 760 g, a little better than the 798 g control but not as good as Liqui Moly or Lubegard | flowed a little faster than motor oil, but viscosity was not too far off from the control | 0.2095 in, a little larger than the 0.207 in control, hardness unchanged at 75 |
| 10Wynn's Engine Stop Leak$24 | got along nicely with the oil, stayed together after 5 minutes with no separation, unlike most other products | 9.35 mm, the largest (worst) wear scar of any product tested; energy use meter in the 580s, test pin extremely hot from metal-on-metal contact | 0.04 g loss, a very low/good result | described as helping restore the O-ring's thickness 'a little', but the specific figure appears to be missing from the transcript at this point (a probable caption drop); hardness given separately | 84.5, softened from the hardened baseline of 89.5 | 686 g, described as showing some improvement over the 798 g control | flowed a little faster than the motor oil control | 0.208 in, described as very little impact, hardness unchanged at 75 |
| 11Lucas Engine Oil Stop Leak$18, tied with Bardahl | very high viscosity, though not quite as high as Bardahl's; no separation from motor oil after 5 minutes | 6.9 mm, a smaller wear scar than the straight motor oil control (7.05 mm) | 0.03 g loss, a very low/good result | 0.197 in, same as XADO | 88, same as Bardahl and XADO | described only as 'a little bit better than the Bardahl, but the improvement is pretty small compared to the Lubegard'; no specific gram figure is given in the transcript for this brand, a probable gap | had a huge impact on cold oil flow performance, grouped with Bardahl as a poor cold-weather result | 0.206 in (identical to the 0.207 in control, essentially no change), hardness unchanged at 75 |
| 12XADO AtomX Stop Leak$20, tied with AT-205 | very low viscosity compared to 5W-30 motor oil, no separation from the oil | 8.8 mm, a large wear scar; a lot of noise and friction, energy use meter over 570 W, test pin pretty hot | 0.06 g loss | 0.197 in, same as Lucas | 88, same as Bardahl and Lucas | 718 g, very close to the Bardahl and Lucas results | moved a little faster than motor oil, but viscosity seemed acceptable | 0.2065 in, described as not hurting the new O-ring, hardness unchanged at 75 |
| 13Bar's Leak Heavy-Duty Formula$8, the cheapest product tested | no separation from motor oil after 5 minutes | 7.52 mm, hurt the performance of the motor oil versus the 7.05 mm control, more friction than straight oil | 0.01 g loss, described as very good (lowest loss of any product tested) | 0.1965 in, very little change from the 0.194 in hardened baseline | remained unchanged at the hardened value of 89.5, no softening at all | 886 g, worse (harder) than the 798 g heat-hardened control, described as unable to soften the gasket | outpaced (flowed faster than) the motor oil control alongside Liqui Moly | 0.207 in and hardness 75, identical to the control, no measurable effect |
| 14Snake Oil | no problems mixing with motor oil, did not separate after 5 minutes | no result found in the transcript; Snake Oil does not appear in the initial lubricity wear-scar test at all, a probable gap (see videoNotes) | only 0.1 g loss, described as performing well in this test | 0.1965 in, did not help restore the O-ring | remained very hard at 89.5, no softening at all | 704 g, performed better than the 798 g heat-hardened control, though not as well as most real stop-leak products | flowed a little too fast, an altered/poor cold result | no result found in the transcript for the 48 hour healthy-O-ring soak test; this control does not appear to have been included in that particular sub-test |
How it was tested
- lubricity/film-strength wear-scar test (50/50 product-oil blend, mm wear scar on a lubricity tester, lower is better, versus a 7.05 mm straight motor oil control)
- evaporative loss and O-ring restoration test: 50/50 product-oil blend plus a pre-hardened O-ring (hardened at 350 F for 2 hours to a baseline 0.194 in thickness / 89.5 hardness) heated to approximately 200 F for 2 hours, then container weight loss (g), O-ring thickness, and O-ring hardness measured
- rubber fiber gasket puncture force test (g force to puncture with a pick) after the same heat exposure, versus a 798 g heat-hardened, untreated control
- cold flow/viscosity test: product-oil blends placed in a freezer at -40 F for 24 hours, then raced down an inclined slide against a motor oil control lane
- new (undamaged) O-ring swell/soften test: healthy O-rings soaked in each product without heat for 48 hours, thickness and hardness measured versus a 0.207 in / 75 hardness motor oil-only control
“If it's all about effectiveness of stopping oil leaks and not about potential risk, the AT-205 easily won this showdown with an average finish of first place.”
Data notes and caveats
13 real stop-leak brands plus a joke/control 'Snake Oil' product (14 total entries), tested across 5 separate sub-tests. Three products (Hapco, Seal Lube, and Snake Oil) are missing a result in the initial lubricity wear-scar test; they were introduced with pricing and mixing-behavior descriptions but no wear-scar mm figure is ever spoken for them, and there is no explicit statement that they were skipped, so this is flagged as a probable transcript/testing gap rather than assumed. This creates an apparent inconsistency in the closing verdict: the transcript states 'when taking into account both the wear scar size and evaporative loss, the AT-205 had an average finish of 3.6 and Seal Lube 4.4', but Seal Lube has no wear-scar figure at all in this transcript, so it is unclear how that average was computed for Seal Lube; both average-finish numbers are reported verbatim without attempting to reconcile the discrepancy. The video gives a separate, broader statement that AT-205 'easily won this showdown with an average finish of first place' (used as the winner verdict here) and separately notes 'all of the top six finishers demonstrated pretty good effectiveness as stop leak products', implying a six-way top tier beyond just AT-205 and Seal Lube; the other four unnamed top-six finishers are not identified in the transcript. Two individual data points are missing/garbled in the transcript rather than the video's own testing gaps: Lucas's gasket-puncture-force gram figure is never spoken (only a qualitative comparison to Bardahl and Lubegard), and Wynn's hardened-O-ring thickness-restoration figure appears to be cut off mid-sentence in the transcript. The narrator closes with an explicit safety caveat distinct from the effectiveness ranking: he would never use a stop-leak product in a vehicle without an active leak, and prefers a proper repair on a newer vehicle, describing his own positive multi-year experience using one on an older vehicle's rack-and-pinion leak. No meta chapters exist for this video.