Retroid Pocket 6 vs G2 (2026): $244 vs a Dead $219

There is a specific comfort in comparing two pieces of hardware side by side. You line up the spec sheets, you find the deltas, you crown a winner, everyone goes home. This comparison denies you that comfort. As of July 2026, one of these two handhelds — the Retroid Pocket G2 — is not a product you can buy. It is a memory. Retroid pulled it from sale on March 16, 2026, a casualty of the same DRAM price spike that has spent the year quietly mugging the entire handheld industry. What follows, then, is half comparison and half autopsy.

The other device, the Retroid Pocket 6, is very much alive and currently selling for an effective $244 after its own mid-life price hike. The G2, when it drew breath, topped out around $219 at retail. That twenty-five-dollar gap is the entire premise the internet wants you to believe this is about. It is not about twenty-five dollars. It is about a former flagship phone chip carrying the best part of a decade of driver optimization on its back, versus a brand-new mid-tier gaming SoC whose software stack was still learning to walk when the plug came out of the wall.

We are going to run the whole gauntlet anyway — specs, silicon, benchmarks from three separate reviewers, battery, a migration guide, five buyer profiles, and a verdict — because the lessons here outlive the G2's short life. The central one, which you already know in your bones if you have ever owned an ARM handheld, is that raw silicon is the easy part and drivers are the hard part. Firmware beats silicon. Let us prove it with numbers.

The $219 Ghost in This Comparison

Before we can compare these two devices honestly, we have to correct the record on how they arrived, because the popular framing — a Pocket 6 from January 2026 squaring off against a G2 from the previous autumn — makes it sound like the G2 is the elder statesman being outgunned by a fresh release. The truth is almost the opposite, and it matters.

A one-day launch gap, not a three-month one

The Retroid Pocket 6 opened pre-orders on October 27, 2025. The Retroid Pocket G2 opened pre-orders on October 28, 2025. That is not a season apart. That is a single calendar day. The "January 2026" date that circulates for the Pocket 6 is when wider retail units actually shipped and reviewers finished their embargoes — a logistics milestone, not a launch. Both machines were designed in the same cycle, announced in the same week, and aimed at overlapping wallets. Treating one as a generational successor to the other is a category error that the spec inversions later in this piece will make embarrassingly clear.

Retroid — the handheld brand under Shenzhen Retroid Technology — has a habit of running parallel product lines rather than a clean tick-tock cadence, and the 6 and the G2 were the clearest example yet. The Pocket 6 was the halo device: bigger, heavier, aimed at people who say the word "Switch" out loud. The G2 was the surgical strike: smaller, lighter, cheaper, and pitched at people who wanted a clean 6th-generation retro machine and no lectures about ray tracing.

The DRAM crunch that killed the G2

Then the memory market detonated. Through 2026, LPDDR5X and UFS pricing spiked as fabs reallocated capacity toward the HBM stacks that AI servers eat by the pallet. Handheld makers, who buy memory in modest volumes and cannot absorb a component doubling in cost, took the hit directly. Retroid took it twice. On March 2, 2026 the company raised the Pocket 6's 8GB model and discontinued its 12GB/256GB configuration outright, telling Android Authority that "the recent surge in memory pricing has reached a level that we are unfortunately unable to absorb."

Two weeks later, on March 16, 2026, the G2 itself was pulled. The listing read "Sold Out" within minutes of the announcement and never came back. Retroid framed it as temporary — a pause driven by "ongoing fluctuations in memory pricing," with the device to return "when market conditions allow." As of this writing in July 2026, it has not returned. The most quietly devastating line came from Retro Handhelds' coverage, where the author noted that "part of the problem with the Pocket G2 is that it never really seemed to fit anywhere in Retroid's lineup" — wedged between the cheaper Pocket 5 and the pricier Pocket 6 with barely any daylight on either side. When something has to be cut, the thing with no clear seat is the thing that goes.

How to read an autopsy

So why compare a living device to a corpse at all? Because thousands of G2 units are in the wild, the second-hand market is active, and the questions the G2 raised — is a newer, cheaper, purpose-built gaming chip better than an older repurposed flagship? does Android version matter for emulation? is raw GPU throughput the number that counts? — are permanent questions that will resurface with every future handheld. The G2 is the cleanest test case we have. If you own one, this piece tells you whether to upgrade. If you are eyeing one used, it tells you what you are buying. And if you simply want a G2-class experience today, Retroid's own answer is the Pocket 5, the device it now steers G2-shoppers toward. Keep that escape hatch in mind; we return to it in the verdict.

Specs Head to Head

Here is the full side-by-side, sixteen rows deep, before we start arguing about what any of it means. Read it once for the obvious gaps — battery, refresh rate, video out — and then read it again for the three rows where the older device somehow ships the newer component. Those inversions are the whole story.

The three spec inversions

Look at the operating system, Bluetooth, and Wi-Fi rows together. The G2 ships Android 15 against the Pocket 6's Android 13. The G2 ships Bluetooth 5.4 against the Pocket 6's 5.3. But the Pocket 6 ships Wi-Fi 7 against the G2's Wi-Fi 6. Two of the three "newer" boxes belong to the cheaper, now-dead device. This is exactly what you would expect from two products built in the same week on different platforms — the G2's newer SoC platform brought a newer Android base-support package and Bluetooth stack, while the Pocket 6's flagship-derived board carried the newer Wi-Fi radio.

Do not mistake any of this for a performance signal. Android 15 does not emulate a GameCube faster than Android 13; the emulator does not care which dessert Google named the OS after. The version numbers are marketing gravity, and if you let a spec-comparison site tell you the G2 is "more modern" because it counts to fifteen, you have been played by a sorting algorithm.

Where the Pocket 6 wins on paper, cleanly

Strip out the inversions and the Pocket 6 takes the rows that actually change your experience. The 120 Hz AMOLED panel is double the G2's 60 Hz — meaningful for front-ends, for high-refresh Android games, and for the general tactile sense that the UI is not wading through syrup. The 6000 mAh battery is a fifth larger than the G2's 5000 mAh cell, and it is fed by 27 W charging. Video out is the blowout: the Pocket 6 drives 4K at 60fps over USB-C 3.1 DisplayPort, while the G2 tops out at 1080p60 over a plain USB-C path. If a dock and a television are anywhere in your plans, this row alone decides it.

Where the G2 wins, and it is not nothing

The G2 is 40 grams lighter and meaningfully smaller in every dimension — 199 mm long against 210, and a slimmer 15.6 mm against 17.2. For a device you hold for three-hour stretches, that is not a rounding error; it is the difference between a machine that disappears into your hands and one you are aware of. Both share the genuinely good stuff — Hall-effect sticks immune to drift, analog L2/R2 triggers, a 3.5 mm jack that refuses to die, and active cooling — so the G2 gives up none of the tactile fundamentals to be the smaller device. If ergonomics and pocketability sit at the top of your list, the corpse has a real argument.

The Silicon: 8 Gen 2 vs G2 Gen 2

Now the part everyone gets wrong. The reflex is to read "Snapdragon 8 Gen 2, released 2022" against "Snapdragon G2 Gen 2, released 2025" and conclude the G2 has the newer, therefore better, chip. Both halves of that reflex are wrong. Let us take the chips one at a time.

Snapdragon 8 Gen 2 (SM8550, "Kalama")

The Pocket 6's brain is the Snapdragon 8 Gen 2, Qualcomm's flagship phone SoC of its year, codename Kalama, part number SM8550. It is built on TSMC's 4 nm N4P process — worth stating plainly, because the brief framing implies only the G2 is a 4 nm part, and that is false; both chips are 4 nm. The 8 Gen 2's CPU is a 1+4+3 cluster: one Cortex-X3 prime core at 3.2 GHz, four Cortex-A715 performance cores at 2.8 GHz, and three Cortex-A510 efficiency cores at 2.0 GHz. The GPU is the Adreno 740, running around 680 MHz here, with full Vulkan 1.3 and hardware ray tracing it will essentially never use in an emulator.

The decisive attribute of this chip is not any single clock. It is that the 8 Gen 2 sat inside millions of Android flagships, which means its GPU driver stack has been sanded smooth by years of real-world abuse — including, critically, the open-source Turnip Vulkan driver from the Mesa project, which the emulation community leans on to fix what the stock Adreno blob gets wrong. A chip with a mature driver is worth more than a faster chip with a raw one. Hold that thought.

The Snapdragon G2 Gen 2 and its Adreno A22

The G2's brain is the Snapdragon G2 Gen 2 — yes, a device called the Pocket G2 running a chip called the G2 Gen 2, a naming coincidence sent to torment search engines. It belongs to Qualcomm's dedicated handheld-gaming G-series rather than the phone line, an eight-core design clocking roughly 1.9 to 2.8 GHz, paired with an Adreno A22 GPU. It is also a 4 nm part, and on raw specifications it is a perfectly respectable mid-tier gaming SoC — that is precisely what it was engineered to be.

The catch is that a purpose-built gaming chip released in 2025 has, by definition, almost no driver history. There were no millions of phones shipping the G2 Gen 2 to shake the bugs out of its GPU stack. When the emulation scene reaches for Turnip on this Adreno A22, it finds a driver that barely knows the silicon exists. That is not a hardware flaw. It is a calendar problem, and calendars are unforgiving.

The raw-power gap is smaller than you think

Here is where the brief's "the Adreno 740 significantly outperforms the A22" narrative falls apart. It does not, at least not on raw throughput. In Retro Handhelds' testing, the G2's single-core CPU score lands about 50% ahead of the Snapdragon 865 that powers the Pocket 5, and only about 10% behind the 8 Gen 2. Its GPU measures roughly double the 865's and, in the reviewer's words, only "8–10% behind the Adreno 740." Read that again: eight to ten percent. In raw silicon, the dead $219 device is within a rounding error of the living $244 one.

Which means the real gap between these machines was never the transistors. Two chips within ~10% of each other on both CPU and GPU do not produce the night-and-day difference reviewers describe in Switch emulation. Something else is doing that work. That something is the driver stack, and it deserves its own section.

The Driver Catch-22

This is the most important section in the article, so I will state its thesis up front and then defend it: the Pocket 6 beats the G2 not because its GPU is faster, but because its GPU is better understood by the software that drives it. On a modern ARM handheld, the driver is the product. The die is just the excuse to ship one.

Turnip, Mesa, and why the Adreno matters more than the numbers

Android's stock Adreno graphics driver is a proprietary Qualcomm blob, and for demanding emulators — Switch above all — it produces glitches: missing geometry, corrupted effects, shader compilation stalls. The community fix is Turnip, the open-source Adreno Vulkan driver maintained under the Mesa project. Turnip has been hammered against the Adreno 740 for years because that GPU lived in flagship phones, so on the Pocket 6 it is a genuinely good driver: it clears the stock glitches without crippling performance. This is the same lesson we keep relearning across the category, that firmware maturity routinely beats a spec-sheet win — the software layer decides what the silicon is actually allowed to do.

On the G2's Adreno A22, Turnip is a stranger. The driver was written and tuned against Adreno GPUs it has seen before; the A22 is new, and Turnip's support for it is immature. So the G2 owner is caught in a genuine catch-22: the stock driver glitches on Switch, and the open-source driver that would fix the glitches tanks the framerate because it does not yet know how to talk to this GPU efficiently. Neither path is good. There is no third path. You wait for the Mesa developers to catch up, on a device that got discontinued before they did.

What simply refuses to run on the G2

The immaturity is not confined to emulators. Because the G2's SoC is so new, chunks of the ordinary Android app ecosystem never got certified for it. HandheldRank, in its head-to-head, is blunt about it: "Some major Android apps straight-up don't work. Netflix games? Nope. Certain big Android games? Nope. Fortnite? Nope." If part of your handheld's appeal was that it is also a competent Android tablet — streaming, storefront games, the occasional battle-royale on the couch — the G2 quietly fails a set of tests the Pocket 6 passes without thinking. The 8 Gen 2 is a known quantity to every app developer alive. The G2 Gen 2 is not.

The verdict the reviewers reached first

This is why the outlets that actually put both devices through Switch emulation did not hedge. HandheldRank's summary of the Switch category is a single sentence: "The RP6 wins here, and it's not close." Their explanation is exactly the driver argument: "The Snapdragon 8 Gen 2 has years of driver optimization from the Android phone ecosystem (Turnip Drivers). The G2's newer GPU lacks that maturity." They land on a framing I will borrow for the rest of this piece — the Pocket 6 is "the safer long-term bet," while the G2 is "the faster way to get a solid handheld." Both halves are true. The G2 was a great way to get an 85%-of-the-way retro machine on day one. It was a poor way to bet on the two years after day one, which is roughly when the Mesa team would have finished the A22 driver — had the device survived that long.

Benchmarks & Battery

Enough theory. Here are the numbers, pulled from three separate reviewers so you are not taking any single site's word for it, plus the battery figures that a spec sheet will never give you honestly.

Synthetic: Geekbench 6 and the CPU gap

On Geekbench 6 single-core, the Pocket 6 scores roughly 1,985. For scale, the Pocket 5's Snapdragon 865 manages about 1,176, so the 8 Gen 2 is a genuine ~69% single-core jump over the previous Retroid flagship — the reason we called the Pocket 6 a roughly 70% CPU uplift for $45 more in the sibling comparison. The G2 lands around 1,780, per Retro Handhelds' "+50% over the 865, −10% versus 8 Gen 2" characterization. So the synthetic CPU gap between the two stars of this article is about 10%. On the GPU side, Retro Handhelds pegs the A22 at roughly double the 865 and 8–10% behind the Adreno 740. Nothing here supports a landslide. On paper, these are siblings.

Per-system emulation ceilings

Where they diverge is in real emulation, and even there the story is subtler than "6 good, G2 bad." Brandon Saltalamacchia's RetroDodo review (a measured 8.4/10) puts the Pocket 6's ceilings at GameCube at 3× native resolution — F-Zero GX and Rogue Squadron running clean — PS2 at 1.5× to 2× with God of War II pushing ~2.5×, Wii genuinely practical (Mario Galaxy, Xenoblade, Donkey Kong Country Returns), and Dreamcast, PSP and PS1 sitting at a comfortable 4×. Switch is "select titles only," gated entirely by the driver situation above. PS3 and Xbox 360 via RPCS3 are a slideshow; the Pocket 6 is a 6th-generation-and-earlier machine that happens to touch the seventh, not a PS3 box.

The G2, meanwhile, is no slouch inside its lane. Noah Kupetsky's Steam Deck HQ review clocks it at PSP 4×, PS2 2.5×, and PS3 "not enjoyable" — which is to say the G2 does PS2 and GameCube-era content, up to native 1080p, without breaking a sweat. That aligns with the general read that the G2 is "fine for PS2 and lower." The delta between the two devices is real but narrow until you reach the exact frontier — Switch — where drivers, not clocks, draw the line.

Battery under actual load

Battery is where the 6000 mAh cell earns its 40 grams. RetroDodo measured the Pocket 6 at around 4.5 hours of mixed use, stretching to 8–10 hours on light 8/16-bit systems (SNES through GBA) and collapsing to 2.5–3 hours when you pin the SoC with full-fat PS2 or GameCube upscaling. The G2, with its smaller 5000 mAh battery and lower-power chip, lands in the same neighborhood for lighter systems but has less headroom for the demanding stuff. If your library skews retro, both last a plane ride; if it skews sixth-gen-upscaled, the Pocket 6's larger cell is the one you want, and its 27 W charging gets you back in the game faster besides. For the tinkerers assembling their own library and front-end, our walkthrough of RetroArch's 200-plus cores is the natural next stop on either device.

Pricing & Availability

This is the section where the comparison stops being fair, because you cannot buy one of the two products at any price. Let us lay out what each cost, what each costs now, and what you can actually put in a cart in July 2026.

The RAM crisis that reshaped the whole table

Every asterisk in that table traces back to memory pricing. The Pocket 6 launched at a $209 pre-order for the 8GB model, settled to a $229 retail price, then jumped to an effective $244 (listed at $249) on March 2, 2026 when Retroid raised it and killed the 12GB/256GB SKU. The 12GB configuration later clawed its way back in June 2026, but only as a reworked 12GB/128GB "stick-top" variant at $279 — same money, half the storage, because storage got expensive too. The G2 walked an identical plank: $199 for the first couple of weeks, $219 at retail, then gone. This is not Retroid being greedy. It is a component shock passing straight through a thin-margin business, and it is the same shock that has been rippling across the entire handheld market all year.

Why the G2 was the one that got cut

Given two devices squeezed by the same crisis, why kill the G2 and keep the 6? Positioning. The G2 sat in the narrow band between the $199 Pocket 5 and the $244 Pocket 6, offering — after the driver caveats — not enough separation from either to justify a third production line during a supply crunch. Retro Handhelds put it precisely: the G2 "never really seemed to fit anywhere in Retroid's lineup." When you have to trim, you trim the SKU whose customers can be absorbed by the neighbors on both sides. G2 shoppers could be pushed down to the Pocket 5 or up to the Pocket 6. The math wrote itself.

What you can actually buy in July 2026

Practically: the Pocket 6 8GB/128GB at $244 is in stock. The 12GB/128GB stick-top at $279 is in stock for the RAM-hungry. The G2 is not in stock, was not restocked, and Retroid's own guidance sends its would-be buyers to the Pocket 5, still hovering around $199–209. Everything downstream of this article's verdict flows from that single availability fact. You are not choosing between a $244 device and a $219 device. You are choosing between a $244 device and a photograph of one.

Five Buyers, Five Verdicts

Spec tables do not buy handhelds; people with specific needs do. Here are five of them, and which device — including the ghost — actually fits.

The Switch-curious power user

You want to play the occasional Switch title, you have realistic expectations about "select titles," and you accept that this is the bleeding edge of ARM emulation. Buy the Pocket 6, no hesitation. This is the exact frontier where the driver-maturity gap becomes a chasm — HandheldRank's "it's not close" was written about you. The G2's Adreno A22 leaves you stuck between a glitching stock driver and a framerate-tanking Turnip build, on a device nobody is optimizing for anymore. The 12GB stick-top variant is the smarter long-run pick here, since Switch emulators are the one workload that genuinely uses the extra RAM.

The PS2 / GameCube purist

Your ceiling is the sixth console generation. God of War II, Gran Turismo 4, F-Zero GX, Wind Waker — you want them upscaled, smooth, and portable, and you have no interest in anything newer. Either device does this beautifully, and this is the buyer the G2 was built for: Kupetsky measured PS2 at 2.5× and PSP at 4× on it, native 1080p without drama. If you find a G2 used at a fair price and this is your whole library, it is a defensible buy. If you are shopping new, the Pocket 6 does the same job with more battery and a 120 Hz screen. And yes, before the inevitable email: emulating hardware you own is settled law — the Ninth Circuit blessed it in Sony v. Connectix, 203 F.3d 596 (9th Cir. 2000), calling the emulator "modestly transformative." The ROMs remain your own conscience's problem.

The pocketability minimalist

You value a device that vanishes into your hands over one that emulates a seventh console you will never touch. The G2 is 40 g lighter and smaller in every dimension, and it keeps the Hall sticks and analog triggers while doing it. This is the one profile where the dead device is genuinely the better object. The tragedy is that you cannot buy it new, and the Pocket 6 at 320 g and 210 mm is a noticeably bigger slab. If small is non-negotiable and you refuse the used market, look at the Pocket 5 or a smaller Retroid entirely rather than forcing the Pocket 6 into a pocket it does not fit.

The couch / TV-out player

You intend to dock this thing and play on a television as often as in your hands. Buy the Pocket 6 and do not think twice. Its USB-C 3.1 DisplayPort output drives 4K at 60fps; the G2 caps at 1080p60 over a plain USB-C path. That is a real, visible difference on a modern panel, and it is one row where the two devices are not siblings at all. Pair it with a proper front-end — some couch players even skip Android entirely for a leanback OS, which is where our guide to flashing Batocera 43.1 comes in handy for the TV-first setup.

The value hunter who missed the window

You wanted the $219 G2, you waited, and now it is gone. Welcome to the largest buyer segment this article serves. Your realistic options are two: pay $244 for the Pocket 6 and get the strictly better long-term device, or drop to the Pocket 5 at ~$199—209, which Retroid explicitly recommends as the G2's replacement and which — at a ~1,176 Geekbench single-core — still clears PS2 and GameCube for most titles. Do not chase a used G2 above ~$180; at $200-plus used you are paying near-new money for an orphaned driver stack and no warranty. The window closed. The Pocket 5 is the graceful exit.

Migration Guide

If you own a G2 and you are moving to a Pocket 6 — or sideways to a Pocket 5 — the good news is that both are Android devices, so nothing here is exotic. The bad news is that Android's scattered save locations mean a careless copy leaves your memory cards behind. Here is the clean way to do it.

Back up saves, states, and configs first

Before you touch the new device, get everything off the old one. The three things that are genuinely irreplaceable are your in-game saves (battery/SRAM and emulator memory cards), your save states, and your per-game configs and control maps. ROMs and BIOS files are re-downloadable; a 40-hour RPG memory card is not. Copy the old device's storage to a PC over USB, or to the microSD card you will physically move across. The important directories on a Retroid Android handheld look like this:

# RetroArch (shared front-end on both devices)
/storage/emulated/0/RetroArch/saves/      # SRAM / battery saves
/storage/emulated/0/RetroArch/states/     # save states
/storage/emulated/0/RetroArch/config/     # per-core + per-game remaps

# Standalone emulators keep their own trees
/storage/emulated/0/Dolphin/GC/           # GameCube memory cards
/storage/emulated/0/Dolphin/Wii/          # Wii NAND / saves
/storage/emulated/0/aethersx2/memcards/   # PS2 memory cards
/storage/emulated/0/PPSSPP/PSP/SAVEDATA/  # PSP saves

Move the data with adb (or a file manager)

If you have a PC and USB debugging enabled, adb is the fastest, least error-prone path. Pull the whole tree from the G2, then push it to the Pocket 6 after installing the same emulators:

# On the OLD device (G2) -- pull to PC
adb pull /storage/emulated/0/RetroArch/saves   ./backup/saves
adb pull /storage/emulated/0/RetroArch/states  ./backup/states
adb pull /storage/emulated/0/RetroArch/config  ./backup/config
adb pull /storage/emulated/0/Dolphin           ./backup/Dolphin
adb pull /storage/emulated/0/aethersx2         ./backup/aethersx2

# On the NEW device (Pocket 6) -- push it back
adb push ./backup/saves   /storage/emulated/0/RetroArch/saves
adb push ./backup/states  /storage/emulated/0/RetroArch/states
adb push ./backup/config  /storage/emulated/0/RetroArch/config
adb push ./backup/Dolphin /storage/emulated/0/Dolphin
adb push ./backup/aethersx2 /storage/emulated/0/aethersx2

No PC? A local file-manager copy to the microSD, then the card into the new device, gets you the same result — just verify the destination paths match, because a mismatched folder name is the single most common reason a migrated save "disappears."

If you can't get a Pocket 6: the Pocket 5 escape hatch

The migration is identical if you are landing on a Pocket 5 instead — same Android tree, same emulators, same save formats. What changes is your expectations: the Pocket 5's Snapdragon 865 is roughly 40% slower single-core than the Pocket 6's 8 Gen 2, so your PS2 upscaling headroom shrinks and Switch is off the table. But your saves carry over byte-for-byte, and RetroArch's save states are forward-compatible so long as you keep the same core versions. Whichever way you jump, back up before you wipe the G2 — a discontinued device is a bad thing to need a do-over on.

Pros & Cons, Tallied

The full ledger for each device, availability included, because in July 2026 availability is a spec.

Retroid Pocket 6

Retroid Pocket G2

The tie-breakers that decide it

Set the pros and cons beside each other and two rows end the argument. The G2's headline con — you cannot buy it — outranks every pro it has. And the Pocket 6's headline pro — a driver stack that reviewers trust with Switch — outranks every con it has, because that maturity is the one thing the G2 could never buy back at any price and the one thing that keeps a handheld relevant past its launch quarter. Everything else is preference. These two are structural.

The Verdict

Time to stop hedging and hand down a recommendation, in the order that actually matters.

If both were sitting on the shelf

In the counterfactual where the DRAM crisis never happened and both devices were in stock at $244 and $219 — the Pocket 6 is still the pick for most buyers, but it would be a real decision rather than a foregone one. The G2 is smaller, lighter, and within ~10% on raw silicon, and for a buyer whose ceiling is PS2 and GameCube, the $25 saving and the 40 grams shed are a legitimate trade. The Pocket 6 wins that hypothetical on drivers, battery, refresh rate, and video out — the safer long-term bet, in HandheldRank's phrase — but the G2 keeps its dignity. It was never a bad device. It was a differently-shaped one.

The reality of July 2026

We do not live in that counterfactual. The G2 has been discontinued since March 16, 2026, was never restocked, and Retroid itself now points its would-be buyers elsewhere. So the practical verdict collapses to something almost anticlimactic: buy the Retroid Pocket 6 at $244, because it is the device you can actually own. If that price stings, the Pocket 5 at ~$199–209 is Retroid's sanctioned G2 replacement and a fine PS2-era machine. The only scenario in which the G2 is the answer is a used listing under about $180 for a buyer who wants a tiny 6th-gen handheld and accepts the orphaned drivers — a narrow slot, honestly reached.

The one-line recommendation

The comparison the internet framed as "$244 versus $219" was never about the money and never really a two-horse race. It was a demonstration that on modern ARM handhelds, a mature driver stack on an older flagship chip beats a raw one on a newer purpose-built chip — and that a device you can buy beats a device you cannot, every single time. The Pocket 6 is not exciting. Saltalamacchia called it "a perfect, yet slightly dull Android handheld," and he is right. But dull and available and well-driven wins this one going away. Buy the living device.