A little more than two years after its release, the Model O Wireless gets a thorough update with the Model O 2 Wireless. Compared to the wired Model O 2, the updates are less numerous and significant. For the most part, the Model O 2 Wireless just gets a visual update, along with a new sensor. Unfortunately, a whole host of performance-related issues make it follow behind even its predecessor.

The most obvious change is the way the Model O 2 Wireless looks. Instead of hexagonal holes, circular ones are now used, and the main buttons no longer extend across the entire width, and instead are framed by ledges at the sides. The aesthetic makeover has little effect on the weight, which is virtually unchanged. At 71 g, the Model O 2 Wireless actually ends up slightly heavier than the Model O Wireless, though in practice, this isn't noticeable. The more worrying part is that Glorious still feels the need to resort to holes for weight-saving at all, as the market has clearly shifted back to solid-shells designs, either with an open or closed bottom. Hence, while the weight itself is fine in my book, having to deal with holes to achieve it isn't. Considering that a design such as the HyperX Pulsefire Haste 2 Wireless manages to get the weight down to 60 g without any holes, though admittedly with less RGB lighting, the Model O 2 Wireless looks wholly unimpressive next to it. Build quality on my sample at least is in a good spot: There is no rattle when shaking the mouse, just minor creaking but no flexing of the shell when applying lateral force, and activating the side buttons by pressing below them is impossible.

The Model O Wireless already utilized the so-called Glorious switches, which in fact are modified Kailh GM 8.0, for the main buttons, and this continues to be the case on the Model O 2 Wireless. Button response thus is expectedly firm and snappy, though definitely on the stiffer side of things, and the revised button design prevents lateral movement, which wasn't always the case on the Model O Wireless. Much like on the wired Model O 2, Huano switches are used for the side buttons, which are a good choice in theory, but the mechanical design isn't particularly good, resulting in noticeable pre-travel on the back button, whereas the forward button is fine. This has the back button feel rather mushy, and the actuation point isn't entirely even, either. The scroll wheel fares much better: An encoder from F-Switch is used, which scores with low noise levels and great tactility, as each step is nicely defined. The feet are unchanged compared to the Model O Wireless both in terms of size and shape as well as material, as pure PTFE is used, and glide is very good still. Likewise, the charging cable is no different, too, and continues to rank among the more flexible charging cables. A rather curious design choice is the anchor, which prevents the cable from being unplugged all too easily during charging, but due to the way it is shaped, also prevents plugging in the cable in either orientation, which is one of the main perks of USB Type-C.

Whereas the Model O Wireless paired PixArt's PAW3370 sensor with a Nordic nRF52840 MCU, the Model O 2 Wireless opts for PixArt's latest PAW3395 in conjunction with a PixArt PAR2862QN, which is very similar to the MCU used for the VAXEE XE Wireless and NP-01S Wireless. Hence, one may expect similarly stellar performance, but this isn't really the case. First of all, CPI deviation isn't just higher than it should be, it also is highly inconsistent. The level of inconsistency exhibited by the Model O 2 Wireless in this regard would otherwise rather be expected from a mouse using a PixArt PAW3335, but not the 3395. Essentially, every step will have varying amounts of positive deviation, which makes it unnecessarily hard to adjust for the deviation. In terms of motion delay, the Model O 2 Wireless does quite well, and musters numbers similar to VAXEE's wireless offerings. In wireless (2.4 GHz) operation, motion delay sits around 0.5 ms, albeit only without MotionSync. MotionSync is a sensor technology used to synchronize SPI reads with USB polling events, improving timing consistency in the process, albeit at the cost of some added motion delay, which in the case of the Model O 2 Wireless, is equal to half of the set polling rate, i.e., 0.5 ms at 1000 Hz. General tracking is mostly fine regardless of MotionSync, with one exception: When moving from a standstill in wireless mode, I've found that after the first few counts, tracking occasionally is suspended momentarily, only to resume normally afterwards. I do not know what could be causing this, but it certainly isn't intended or good behavior. I've also been able to identify what I like to call "shift" behavior, whereby at some random point during the motion, motion delay will be increased by a single interval, which is something I've previously encountered on many releases, including those from VAXEE and Razer. A more impactful issue is the lack of polling stability on the Model O 2 Wireless. In wired operation, only 1000 Hz is stable, whereas every other polling rate suffers from periodic off-period polls. In wireless operation, the instability persists, but 1000 Hz is likewise affected, and it only gets worse once RGB lighting is added to the mix. In short, there is not a single fully stable polling rate or setting when using the Model O 2 Wireless wirelessly.

When it comes to click latency, the bad news doesn't stop. In theory, the Model O 2 Wireless musters formidable numbers. Once the unacceptably high default debounce time of 10 ms is lowered to 0 ms, click latency sits at 2.3 ms in wireless operation, which is more than fine. However, this number doesn't tell the whole story. Somewhat similarly to mice with the CX52850 MCU, the Model O 2 Wireless suffers from so-called idle delay, which describes additional delay present if there hasn't been any sensor or button activity for a given period of time. On the CX52850, the threshold for when this triggers was five seconds, rendering instances of when this could happen rare edge cases. On the Model O 2 Wireless, however, a mere 0.3 s of not moving the mouse or clicking a button is enough to cause the next click be delayed by varying amounts of time, though typically upwards of 4 ms when taking the click latency value gathered for 0 ms debounce time as the baseline. Hence, the 2.3 ms is a "best case" scenario if anything, and actual click latency will be significantly higher most of the time. The main issue with this isn't so much the absolute latency but rather the degree of inconsistency introduced by it, as the degree by which a click will be delayed can vary by up to 10 ms or more, essentially being completely random. Due to this inconsistency, I consider the Model O 2 Wireless not suitable for any level of competitive play.

The purpose of this idle delay is likely saving power, which leads to another issue. Glorious indicates a battery life of up to 110 hours in 2.4 GHz wireless operation without illumination, and up to 210 hours using Bluetooth. These numbers sound great on paper, but of course, on an RGB-laden mouse such as the Model O 2 Wireless, knowing the battery life with RGB enabled would be worthwhile. Determining that proved difficult in my testing. After using the Model O 2 Wireless continuously for 8 hours with RGB lighting enabled (default brightness), the indicator still stood at 100%, which no doubt is a buggy reading. At a later point, I've had the mouse lying flipped on the desk for a little less than two hours, which prevented it from shutting automatically off, and the indicator then showed 42%, down from 100%. Clearly, losing more half its charge after two hours of being turned on isn't a real possibility. To sum it up, battery readings are as useless as they can be, turning any estimations into a guessing game. For the record, a 450 mAh battery is utilized, so battery life shouldn't be catastrophic at least. The battery charges at a little less than 0.3 A, which is decently speedy.

Battery reading randomness aside, Glorious Core, which is the software for the Model O 2 Wireless, works decently well, though resource usage continues to be significant. A recent update has lowered RAM usage to more reasonable levels at least, but CPU usage is still higher than it should be, at least when the software isn't minimized. Compared to the likes of Razer Synapse, Corsair iCUE, or ASUS Armoury Crate, one laudable feature of Core is the fact that no processes keep running after exiting the software.

Whereas the Model O Wireless retailed for a highly competitive $79.99, Glorious charges a rather hefty premium of $20 for the Model O 2 Wireless. While $99.99 isn't out of line with many competing products, considering the numerous issues the Model O 2 Wireless suffers from, it becomes a tough sell nonetheless. For the same money, one could get the LAMZU Atlantis 4K, which pulverizes the Model O 2 Wireless on pretty much every level, whereas on the lower end, the Darmoshark M3 likewise is the better mouse, for half the money. In fact, I'd even opt for the original Model O Wireless over the Model O 2 Wireless, as it performs better while being more affordable. In short, making a case for the Model O 2 Wireless is most difficult, at least for the time being.