Gaming mice exclusively made for fingertip grip are still a rather rare sight—a niche within a niche, if you will. As such, it comes as no surprise that the few that exist typically cater to enthusiasts, especially when it comes to price. The Zaunkoenig M2K at $300 was one such an example, and the G-Wolves HSK Pro at $179.99 isn't exactly a bargain, either. At $89.99, the Darmoshark M5 may cost more than your average Darmoshark mouse, but still remains much more affordable than the aforementioned releases.

The HSK Pro is a good reference point, as the M5 essentially has the same shape, albeit uniformly scaled-up. Accordingly, the M5 likewise has a large hump in the back, lending itself at least passably well to claw grip, although for small hands only. At 38 g, the M5 weighs a good 12 g more than the HSK Pro, which is only partially due to the M5 being larger, not having holes, and having virtually twice the battery capacity. Unlike on the HSK Pro, where the hump had to be there to make room for side buttons, this isn't the case on the M5, which just as well could've done with less hump and therefore less volume, and the PCB on the M5 is much more spacious than on the HSK Pro, and more than it would need to be in general. In addition, Darmoshark has also added superfluous things such as an LED on the PCB, which may add less than a gram of weight, but is utterly useless in providing scroll wheel lighting or serving any kind of other purpose. At 38 g, the M5 sure isn't heavy, but for a fingertip grip mouse, every gram counts, and the shape is basically irrelevant beyond the finger contact points. The M5 does have flat sides at least, and being larger than the HSK Pro, it'll also feel less cramp-inducing due to being wider, and almost as wide as the M2K. As far as build quality goes, I don't have much to complain about on my sample. There is no rattle when shaking, no creaking or flexing when applying lateral pressure, and while the side buttons can be actuated by pressing below them, doing so requires excessive force. The only flaw I could find is with the right main button on my copy, as when pressed right at the front, button response will often be muted and hollow. Due to the way the mouse is designed, I'd argue having one's fingers right at the front is the default, so this deficiency is felt all the time. I haven't been able to establish where this issue originates precisely, though I also believe that this isn't a common occurrence, and on average shouldn't be an issue. Still, something to keep in mind nevertheless.

Speaking of the main buttons, these are outfitted with Kailh switches, more specifically of the white transparent shell, white plunger variety, also known as "White Sword." Button response is firm and snappy, despite pre and post-travel being moderate, and moderate amounts of lateral movement being present upon provoking it. For the side buttons, surface-mounted Kailh switches are used, which comes as no surprise given how small these buttons are. For their size, button feedback is alright, and both pre and post-travel are pleasantly low. The scroll wheel encoder comes from F-Switch (yellow, green core), and while tactility is above average, with the individual steps being rather well-separated from each other, noise levels are very substantial regardless of scroll direction, and the wheel feels a bit flimsy as well. The feet are made of white-dyed pure PTFE, whose glide is fine, and a set of replacement feet is found in the box, along with a set of grip tape.

Darmoshark pairs PixArt's PAW3395 with a Nordic nRF52840 MCU on the M5. Originally launched with support for polling rates of up to 4000 Hz in wireless operation, support for 8000 Hz has been added through a firmware update. Using the latest firmware, performance is excellent in some areas and spotty in others. Right out of the gate, CPI deviation is an issue on the M5. On most 16-pin sensors such as the 3395, CPI deviation can be of any degree, but this degree will be consistent across the entire CPI range, barring possible issues with nominal CPI steps being mislabeled. On the M5, however, deviation not only is high, but all over the place too, essentially forcing one to measure it oneself to be able to compensate for it. If anything, this at least qualifies as a major inconvenience. General tracking is fine on the M5, regardless of whether MotionSync is enabled or not, and polling rate stability is also in a good place. Both 500 and 4000 Hz show elevated variance in wireless operation, but the target intervals are always met, even at 8000 Hz, and even when enabling MotionSync at 8000 Hz. Motion delay also is highly competitive on the M5. Using the default sensor run mode, the M5 achieves parity with the Logitech G403 (control subject) at 1000 Hz in wireless mode, and by setting the sensor run mode to corded mode, which is done by enabling a so-called "esports mode," the M5 will even be ahead of the G403 by 1.0 ms. At 2000 Hz, this latter advantage is maintained, whereas at 4000 Hz, motion delay constantly varies due to the polling variance mentioned earlier. 8000 Hz, however, can be classified as at least partially broken, as motion delay seemingly increases throughout the motion, and tracking shows a lot of SPI timing jitter as well. In short, sticking to either 2000 Hz or 1000 Hz with esports mode enabled is recommended on the M5, in which case sensor performance will be on a great level. Click latency also looks good: in wireless operation at 1000 Hz, 0.8 ms are averaged, which is lowered to 0.6 ms at 2000 Hz and 0.4 ms at 4000/8000 Hz. These numbers are easily on par with releases from Razer, LAMZU, or Finalmouse, and only the CHERRY XTRFY M64/68 Pro and G-Wolves Hati-R 8K, both of which average 0.3 ms, manage to edge them. That said, the M5 only musters click latency this low if set to the lowest debounce time value of 0 ms, and any values above that will scale linearly. At 0 ms, slam-clicking, which describes inadvertent button actuation upon resetting the mouse, will be quite common, though setting debounce time to 1 ms already sufficed to eliminate it on my unit.

Darmoshark doesn't provide any battery life numbers, though at least a battery life indicator has found its way into the software. Despite being percentage-based with single-digit accuracy, this indicator is far from reliable, as the given value randomly goes up or down, thus making it impossible for me to gauge battery life. At the same time, based on the capacity of 300 mAh as well as the components and firmware used, I consider 70–80 hours at 1000 Hz (esports mode disabled) a plausible range on the M5. In any case, Bluetooth being present is a nice bonus, and should allow one to squeeze at least 100 hours out of a single charge. Unfortunately, charging isn't exactly speedy at less than 0.2 A, and the charging cable is both too stiff and too short (1.50 m) to comfortably allow playing while charging. Accordingly, keeping an eye on charge is definitely recommended on the M5. Thankfully, keeping the software installed is not a burden, as resource usage is pleasantly low, and all the important settings are included, with no bugs to speak of.

Overall, I wouldn't say that the M5 would be among my top picks for a fingertip grip mouse, simply by virtue of being relatively heavy for its size, despite its weight being low in an absolute sense. At the same time, at $89.99 on Mechkeys.com, the M5 is easily among the most affordable fingertip grip only mice available, and can serve as a good starting point for those wanting to venture into this grip style without having to pay upwards of $150. In addition, performance is solid overall, and click latency ranks among the very best. With that as well as possible build quality woes in mind, the M5 earns our Recommended award.