The Best Graphics Cards in 2025

View all RTX 5090 cards 

Reasons to Buy:

• Unrivalled gaming performance (fastest GPU ever) – 4K/8K capable
• Massive 32GB VRAM and huge bandwidth for future-proofing
• Excellent ray tracing and AI upscaling (DLSS 4) capabilities
• CCL Advantage: We restock the 5090 regularly – grab the flagship when it’s available!

Reasons to Avoid

• Extremely expensive and often out of stock (high demand)
• Very high power draw (575W) – requires strong PSU and cooling
• Overkill for anything below 4K (bottlenecked at 1080p/1440p in many games)

Nvidia GeForce RTX 5080

The RTX 5080 is Nvidia’s second-in-command, offering flagship-class performance with a slightly more reasonable footprint. It features 84 SMs (10,752 CUDA cores) and 16 GB of GDDR7 memory on a 256-bit bus. Essentially, the 5080 is a cut-down 5090 – it keeps the same architectural features (DLSS 4, Frame Generation, RTX improvements) but with fewer cores and a much lower power rating of ~360 W. In practice, the RTX 5080 handles 4K gaming superbly: you can expect around 70–80 FPS at 4K Ultra in demanding titles (a bit lower with maxed-out ray tracing, unless DLSS is used), and well into the triple digits at 1440p. It’s roughly 10–15% faster than the last-gen RTX 4080 in traditional raster performance, which means it can comfortably drive a 4K 120 Hz display in many games with settings tweaks. In fact, at 4K with heavy GPU load, the 5080 isn’t far off the 5090 – it’s at 4K where it reaches ~78% of the 5090’s performance. The gap is more evident at lower resolutions where the 5090’s extra cores flex. Still, 4K60 gaming is a slam dunk on the RTX 5080, and 4K120 is achievable in plenty of games with DLSS or settings tuning. For 1440p high-refresh (144 Hz+), the 5080 is a fantastic choice, often hitting 120–160 FPS in competitive titles.

The benefits of the RTX 5080 include a more palatable power draw and thermals. At ~360W TBP, it’s significantly easier to cool than the RTX 5090 (many custom cards are dual-slot), and it typically uses just 2×8-pin (via adaptor to the 16-pin) for power. It’s still a high-end card, so you’ll want a quality PSU (750W+). The 16 GB VRAM is plenty for today’s games at 4K, though a few extreme cases might push near that limit with max textures – so far 16 GB has proven sufficient in this class. One thing to note is that the generational leap here is modest; reviewers called the 5080 more of an “incremental upgrade” over the 4080. A lot of its “wow” factor comes from DLSS 4’s frame generation, which Nvidia heavily advertises. Without DLSS 4, the raw improvement isn’t massive (in fact, drivers at launch seemed tuned for DLSS 4 over raw fps). But given that most big games now support DLSS, the 5080 can leverage those features to pull ahead.

Take a look at a few of the RTX 5080 models we currently stock below:

View all RTX 5080 cards

Reasons to Buy:

• High-end performance for $<flagship price> – great for 4K and 1440p high refresh
• More energy-efficient and runs cooler/quieter than the RTX 5090 (360W vs 575W)
• 16GB VRAM is ample for 4K gaming today (and more than competing cards in this range)
• Supports all latest Nvidia features (DLSS 4, Reflex, RTX IO) for longevity

Reasons to Avoid:

• Still expensive (top-tier pricing, though less than the 5090)
• Only a moderate jump from last-gen in raw fps (reliance on frame generation to shine)
• Availability is limited – sells out quickly, albeit a bit easier to find than 5090
• Faces new competition from AMD’s top cards at a lower price point (if you don’t need DLSS)

Mid-Range Graphics Cards: 1440p Powerhouses

Moving down the stack, the mid-range GPUs of 2025 offer fantastic performance for 1440p gaming and strong 4K value. This category is highly competitive, with Nvidia’s RTX 5070 series going head-to-head against AMD’s RX 9070 series. These cards are priced more reasonably and are very attractive to high-volume gamers – expect 100+ FPS at 1440p, and even 4K 60+ FPS with some settings tweaks or upscaling. Interestingly, thanks to DLSS 4, even mid-range cards this generation can sometimes match last-gen flagships in supported. Here are the top mid-range picks:

Nvidia GeForce RTX 5070 Ti

The RTX 5070 Ti is a new entry this generation, sitting between the 5080 and 5070 and providing high-end performance at a slightly lower cost than the 5080. It uses a cut-down GB203 GPU (same chip as the 5080) with 8960 CUDA cores active and retains 16 GB of GDDR7 VRAM on a 256-bit bus. Essentially, the 5070 Ti gives you 5080-like memory and features, but with ~20% fewer shaders and a lower power target (~300W TBP). In gaming benchmarks, it delivers an excellent 100+ FPS at 1440p Ultra (around 101–102 FPS on average across games), which makes it a perfect card for a 1440p 144Hz monitor. Even at 4K Ultra, the RTX 5070 Ti holds up well, averaging around 60 FPS – just shy of the mythical 4K60 baseline in the very heaviest games, but easily surpassing it if you enable DLSS or dial back one or two ultra settings. In fact, the 5070 Ti is only about 4% faster than AMD’s RX 9070 XT and about 18–20% faster than the non-Ti 5070 in raw performance. It roughly matches or slightly exceeds the previous-gen RTX 4070 Ti (and “4070 Ti Super”) it replaces. This means it can even trade blows with an older RTX 4080 in some scenarios, especially when frame generation is in play.

Like its bigger siblings, the RTX 5070 Ti supports DLSS 4 and Nvidia’s full RTX feature set, so it benefits greatly in games that implement DLSS 4’s Multi-Frame Generation. It’s a “proper high-end GPU” for those who don’t want to go all the way to an 5080/5090. The card’s 16GB VRAM is a strong selling point at this tier, especially since 12GB is starting to look borderline for some new titles (we’ll discuss VRAM more in the features section). The RTX 5070 Ti’s main drawbacks are its generational uplift – it’s not a huge leap beyond a 4070 Ti unless you count the DLSS 4 “marketing-heavy multi-frame generation” features. If you can snag one, though, it’s one of the best 1440p GPUs you can get.

Take a look at a few of the RTX 5070 Ti models we currently stock below:

View all RTX 5070 Ti cards

Reasons to Buy:

• Excellent 1440p performance (100+ FPS at Ultra) – a 1440p gamer’s dream
• Capable of 4K gaming with upscaling (around 60 FPS at 4K Ultra native)
• 16GB VRAM provides breathing room for high-res textures and future games
• New Nvidia features (DLSS 4 Frame Generation, Reflex, RTX Video Super Res, etc.) all supported

Reasons to Avoid:

• Only a minor step up from last-gen 4070 Ti in rasterized games
• Still on the pricey side of mid-range
• Multi-Frame Generation claims can be misleading (don’t expect miracles without some latency trade-off)

Nvidia GeForce RTX 5070

Nvidia’s RTX 5070 might be the sweet spot for many gamers in 2025. The RTX 5070 uses a smaller GB205 GPU with 6144 CUDA cores, paired with 12 GB of GDDR7 VRAM on a 192-bit bus. Its power draw is around 250 W, making it a very power-efficient card given its performance. In gaming benchmarks, the RTX 5070 proves its worth: it’s about 20% faster than the prior generation RTX 4070 on average, which places it roughly on par with an RTX 4070 Ti (or “4070 Super”) in many games. In practical terms, the 5070 can easily handle 1440p Ultra at 80–100 FPS in modern games, and even approaches 120 FPS in slightly lighter titles or with DLSS. At 1080p, it’s a high-refresh beast (often 120–160 FPS), though 1080p is not this card’s primary target. It’s also surprisingly competent at 4K: expect roughly 50–55 FPS at 4K Ultra as a baseline (a bit higher in some games, or if DLSS Quality is used). That means with upscaling or a few tweaks, the RTX 5070 can achieve 4K60 gameplay, something mid-range cards of previous generations struggled with.

One of the most interesting aspects of the RTX 5070 is how it performs in DLSS 4 supported titles. Nvidia’s CEO Jensen Huang famously claimed, “The RTX 5070 – 4090 performance, thanks to DLSS 4. This bold claim sparked debate, but there is a nugget of truth: in games that fully utilise DLSS 4’s frame generation, the 5070 can indeed produce frame rates similar to an RTX 4090 running without frame generation in those same games. For example, a 5070 with DLSS 4 MFG might hit ~120 FPS where a 4090 with traditional rendering also hits ~120 FPS. This does not mean the 5070 is as fast in all scenarios (it’s not – raster and ray tracing raw performance are much lower), but it’s a testament to how far AI upscaling has come. The flip side is that those “4090-level” frame rates come with higher latency “fake” frames, so it’s great for visual smoothness but not equal in responsiveness. Nonetheless, for a mid-range card to even be mentioned in the same sentence as a 4090 is remarkable – and it means future-proofing is strong if developers keep integrating DLSS 4/FrameGen.

The RTX 5070 has 12GB of VRAM, which has stirred some conversation. 12GB is enough for 1440p in virtually all current games, but there are edge cases at 4K (with max settings in VRAM-heavy games) where 12GB could become a limit. AMD’s competing cards offer 16GB at this tier, and indeed the 16GB RX 9070 is only about 7% faster than the 12GB RTX 5070 in 1440p performance – a small raw gain, though the extra memory might prove useful over a longer horizon. We’ll talk more about VRAM later, but keep in mind 12GB is okay for now, just not lavish. On the plus side, the RTX 5070 is relatively easy to cool and fit – many models are dual-slot and under 11 inches, and you won’t need more than a decent 600–650W PSU. It launched in February 2025 and saw high demand, but stock has been somewhat better than the higher-end cards. This makes the RTX 5070 one of the best value buys in the current market, hitting that performance-per-dollar sweet spot for high-quality 1440p gaming.

Take a look at a few of the RTX 5070 models we currently stock below:

View all RTX 5070 cards

Reasons to Buy:

• Outstanding 1440p performance for the money – ~20% faster than RTX 4070 for same price
• Can leverage DLSS 4 to achieve near-4090 level FPS in supported games
• Efficient 250W power draw; runs cool and fits in standard mid-tower cases easily
• Most affordable way to get Nvidia’s latest architecture and features (Frame Generation, RTX IO, etc.)

Reasons to Avoid:

• “Only” 12GB VRAM – fine for now, but less than AMD’s alternatives (could affect future ultra settings)
• Raw performance roughly ties RX 9070 (16GB) which some might prefer for the extra memory
• Nvidia’s frame generation isn’t universally supported in all games (raster performance is merely good, not magic)

AMD Radeon RX 9070 XT

AMD’s top mid-range contender is the Radeon RX 9070 XT, and it’s a formidable card in its own right. Built on the new RDNA 4 architecture, the RX 9070 XT features 64 Compute Units (CUs) (versus 96 CUs in a hypothetical high-end RDNA4) and comes with 16 GB of GDDR6 VRAM on (reportedly) a 256-bit interface, plus the latest iteration of Infinity Cache. Its typical board power is around 280 W. In terms of performance, the RX 9070 XT is designed to go head-to-head with Nvidia’s RTX 5070 Ti. In pure rasterized rendering (standard gaming without upscaling), the 9070 XT actually edges out the 5070 Ti by a hair. Internal testing showed the 9070 XT beating the RTX 5070 (non-Ti) by ~24% on average, and coming within 4% of the higher-priced RTX 5070 Ti. In fact, AMD claimed (and reviews echoed) that the RX 9070 XT’s raster performance is roughly on par with last generation’s RX 7900 XTX flagship. In practical terms, this means the 9070 XT can easily drive 1440p Ultra at 100+ FPS and handles 4K gaming around 60 FPS (similar to the 5070 Ti). Both the 9070 XT and 9070 (below) come with 16GB VRAM, which gives them an advantage in memory-heavy scenarios; you won’t need to worry about textures not fitting in VRAM at 1440p or 4K. AMD positioned the 9070 XT squarely as an option for those who want strong raw performance and extra memory for the long term.

Where the RX 9070 XT lags a bit is in ray tracing and AI upscaling (areas Nvidia has traditionally led). The RDNA 4 architecture brought improvements, and indeed the 9070 XT’s ray tracing is better than the previous gen – AMD narrowed the gap such that in many ray-traced games at 1440p, the 9070 XT is close to the RTX 5070 Ti. However, in heavy ray tracing or path-traced games, Nvidia’s cards (with their larger RT core counts and DLSS) still pull ahead. AMD has its own upscaling tech, FSR 4, which now includes AI enhancements and even frame generation on RX 9000-series cards. FSR 4 works, and its image quality has improved to be closer to DLSS, but DLSS 4 on Nvidia still generally provides slightly better results and is available in more titles. So, if you plan to use a lot of ray tracing or are excited about frame generation, Nvidia holds an edge. If raster performance per dollar and more VRAM is your priority, the RX 9070 XT shines. CCL carries custom models from partners like Sapphire, XFX, PowerColor, and ASUS, so there are plenty of choices in cooling and form factor.

Take a look at a few of the RX 9070 XT models we currently stock below:

View all RX 9070 XT cards 

Reasons to Buy:

• Matches or beats the RTX 5070 Ti in standard performance (excellent 1440p/4K speeds)
• 16GB GDDR6 VRAM – more memory headroom than any Nvidia card in this class
• Improved ray tracing cores and new FSR 4 with frame generation support (works on all GPUs, though best on RX 9000 series)
• Typically a bit cheaper than Nvidia’s equivalent, for very similar results (great value when priced right)

Reasons to Avoid:

• Ray tracing performance still lags Nvidia slightly in the most intensive scenarios
• FSR 4 adoption is growing but DLSS still has wider game support and slightly better quality in some cases
• Driver updates – AMD has improved, but Nvidia’s ecosystem (ShadowPlay, NVENC, etc.) is a bit more mature if those matter to you

AMD Radeon RX 9070

Next is the Radeon RX 9070 (non-XT), which is essentially a step down from the 9070 XT. It uses 56 CUs (versus 64 in the XT) and the same 16GB GDDR6 memory. With slightly fewer cores and a lower power draw (~215 W typical, the RX 9070 slots in as a competitor to Nvidia’s RTX 5070 (non-Ti). How do they compare? In short, it’s a close fight. The RX 9070 in our testing is about 7% faster than the RTX 5070 at 1440p and roughly 3–8% faster at 1080p – essentially a negligible difference in many real-world cases. That means the RX 9070 delivers superb 1440p performance, comfortably hitting ~90 FPS averages at 1440p Ultra (just a tad above the RTX 5070’s ~80–85 FPS). At 4K, the gap shrinks; the RTX 5070 might be within a couple frames per second of the RX 9070 since both hover in the mid-50s FPS at 4K Ultra. The extra 4GB of VRAM on the RX 9070 is a nice bonus when running 4K, as it can mitigate any potential texture streaming issues that a 12GB card could face down the line. In fact, if you value memory, the RX 9070 is basically offering RTX 3080 Ti levels of performance but with 16GB instead of 12GB, which is compelling for future-proofing.

Like its XT sibling, the RX 9070 supports FSR 4 upscaling and frame generation, but note that unlike DLSS, FSR can work on any GPU (though FSR 4’s latest features only enable on RX 9000 cards). This means even if a game supports only FSR, an Nvidia card could use it too (albeit without hardware AI acceleration). The RX 9070’s ray tracing is a step up from last-gen AMD (it outperforms the old RX 7900 XT in ray tracing, for example), yet when compared to Nvidia’s RTX 5070, it’s roughly a wash – Nvidia is a bit faster in heavy RT effects, but not by a huge margin this generation. Both cards can do ray tracing at 1440p pretty well, especially with upscaling. Also, AMD’s cards offer features like Smart Access Memory (if you have an AMD CPU) to get a few extra percent performance. CCL has models from Gigabyte, Sapphire, XFX, etc., and the RX 9070 has been a bit easier to find in stock than some Nvidia cards (possibly because initial demand gravitated to the XT).

Take a look at a few of the RX 9070 models we currently stock below:

View all RX 9070 cards

Reasons to Buy:

• Great performance for 1440p gaming (neck-and-neck with RTX 5070, slightly faster in many titles)
• 16GB VRAM – best-in-class memory for a “mainstream” GPU, ideal for longevity
• Efficient for its performance (sub-220W usage, many cards are compact dual-fan designs)

Reasons to Avoid:

• Only modestly faster than the RTX 5070 in most scenarios (single-digit % differences)
• Relies on FSR for AI upscaling – which, while improved, is still catching up to DLSS 4 in quality and support
• Ray tracing improvements are solid, but still a notch below Nvidia’s best in class (especially at 4K with heavy RT)
• AMD’s driver feature set (ReLive, etc.) is good, but Nvidia’s ecosystem might have features some creators prefer (if that matters to you)

Budget Graphics Card Options

Not everyone needs a top-of-the-line card – especially if you game at 1080p or are building a PC on a tighter budget. The good news is that 2025’s budget and lower-midrange GPUs are quite capable, often matching the high-end cards from a few years ago. In this segment, Nvidia has filled out the lower end of the RTX 50-series with the RTX 5060 Ti (in 8GB and 16GB variants), with an RTX 5060 (non-Ti) expected to launch soon. AMD’s equivalent RX 9000-series budget cards (like an RX 9050 or 9060) haven’t been announced yet, so the focus here will be on Nvidia’s offerings which are available now. These cards are ideal for 1080p high settings gaming at 60–144 FPS, and even 1440p with some compromises or the help of upscaling. Importantly, Nvidia addressed a critical concern at this tier: VRAM. The RTX 5060 Ti comes with a choice of 8GB or 16GB, and as we’ll explain, that can make a big difference for longevity.

Nvidia GeForce RTX 5060 Ti (16GB)

Nvidia’s RTX 5060 Ti 16GB is arguably the most well-rounded budget GPU in 2025. It’s built to succeed the RTX 4060 Ti, but Nvidia learnt from the feedback on that last 8GB card and released a 16GB model right off the bat (alongside an 8GB model). The specs: the 5060 Ti uses a cut-down GPU (likely a variant of GB205/206) with performance around ~15% higher than the RTX 4060 Ti. Reviews indicated it’s 10–20% faster than the 4060 Ti in gaming tests, putting it in the performance ballpark of an RTX 3070 or slightly above. At 1080p, the RTX 5060 Ti 16GB absolutely flies – you can expect well over 100 FPS in most games at high or max settings. For instance, in one benchmark (Final Fantasy XIV), the 5060 Ti hit 104 FPS at 1440p, indicating how strong it is at 1080p (CPU limits often kick in beyond ~150 FPS at 1080p). At 1440p, this card holds up impressively: in more GPU-intensive titles, it averages around 55–70 FPS at 1440p Ultra. In fact, one report showed the 5060 Ti achieving ~57 FPS at 1440p in a test, placing it right between AMD’s RX 7800 XT and Nvidia’s RTX 4070 in that scenario – not bad for a “60 Ti” class card! 4K gaming is a stretch for the 5060 Ti (expect maybe ~30 FPS at 4K Ultra), but with DLSS 4 upscaling, even 4K 60 can be within reach in older or lightweight games. The key advantage of this model is that 16GB VRAM. At 1080p, 16GB means you’ll never worry about memory limits, and at 1440p it provides safety for future games. Titles like Hogwarts Legacy, The Last of Us, etc., which on 8GB cards had to drop texture quality to avoid slowdowns, will run smoothly on the 16GB 5060 Ti. Basically, this card addresses the longevity concerns – it’s a midrange performer with high-end memory capacity for its class.

Power-wise, the RTX 5060 Ti 16GB is rated around 150–180W (exact TBP TBC, but it’s low), which makes it super efficient. Most models don’t even need more than a single 8-pin power connector. It runs cool and quiet; small dual-fan coolers are sufficient. This makes it great for small form factor builds or upgrading older OEM PCs. Reasons to choose the 16GB over the 8GB variant (below) are straightforward: if you plan to keep the card for several years or play new AAA titles, the 16GB ensures you won’t hit a VRAM wall at high settings. The cost is slightly higher, but many reviewers and gamers consider the 16GB version the “only option” at this point for a truly future-proof 1440p experience. CCL offers multiple brands (ASUS, Zotac, MSI, Palit, etc.), and we strongly recommend the 16GB model for anyone shopping in this tier unless your budget is extremely tight.

Take a look at a few of the RTX 5060 Ti 16GB models we currently stock below:

View all RTX 5060 Ti cards

Reasons to Buy:

• Best-in-class VRAM: 16GB on a budget GPU is excellent for new games
• Great 1080p performance (matches or exceeds RTX 3070 levels)
• Strong 1440p capability for the price – ~60+ FPS in demanding games, even better with DLSS
• Low power, cool running, and compact – fits in almost any build easily

Reasons to Avoid:

• Slightly more expensive than the 8GB variant (but for good reason)
• Not intended for 4K gaming (4K requires low settings or DLSS Performance mode)
• Faces no direct AMD competition yet, but AMD could release something in this range (if you’re not in a hurry, you could wait to see)

Nvidia GeForce RTX 5060 Ti (8GB)

The RTX 5060 Ti 8GB is essentially the same GPU as the 16GB version, but with half the VRAM. It’s an interesting option: in pure performance, the 8GB card performs nearly identical to the 16GB in most current games (within a couple percent), since they have the same core config and bandwidth. For example, at 1080p and 1440p in many titles, you’ll see the RTX 5060 Ti 8GB achieve the same frame rates as the 16GB. However, the difference emerges in memory-intensive scenarios. In some modern games, especially at 1440p Ultra textures, the 8GB buffer can lead to lower minimum framerates or stutters when it runs out of memory. We’ve already seen cases where the 16GB card “delivered slightly better performance” simply because the 8GB card had to drop texture quality or suffered minor hitching due to memory swapping. As time goes on, such scenarios might become more common. That said, if you only play esports or older titles, or you stick to 1080p medium/high settings, 8GB will serve you just fine. It’s also worth noting that DLSS 4’s frame generation can somewhat mitigate heavy GPU load – both the 8GB and 16GB benefit from DLSS, but if a game needs more than 8GB at certain settings, DLSS won’t fully solve that.

The RTX 5060 Ti 8GB’s main appeal is cost savings. If you are on a strict budget and every pound counts, the 8GB can save you some money while still giving you the same shader performance. It’s a good card for a 1080p 144Hz focus where you might play competitive games that aren’t VRAM hogs. However, many in the community (and we at CCL) lean towards recommending the 16GB variant if you can afford it, simply to “future-proof” your purchase. One telling quote summed it up: “c’mon, you know the 16GB is the only option if you want to play the latest and greatest games at 1440p now and in the near future.”. That said, if your use case is firmly 1080p and you upgrade GPUs every couple of years anyway, the 8GB 5060 Ti is still a very capable card that outperforms anything else at its price point from the previous gen (it handily beats the old RTX 4060 Ti and trades blows with a 4070). Just be mindful of the memory limitation – maybe avoid maxing texture packs in certain games to stay within that 8GB comfort zone.

Take a look at a few of the RTX 5060 Ti 8GB models we currently stock below:

View all RTX 5060 Ti cards

Reasons to Buy:

• Same great raster performance as the 16GB model in most games (outperforms last-gen equivalents)
• More affordable – can save money if 16GB’s price premium is too much for your budget
• Ideal for 1080p gaming or 1440p with medium settings (8GB covers those scenarios well)
• All the RTX 50-series features (DLSS 4, etc.) at a lower price entry point

Reasons to Avoid:

• 8GB VRAM may struggle with future games at Ultra settings (already a limit in a few recent titles)
• Resale or longevity might be weaker due to lower VRAM (potentially less “proofed” for the future than the 16GB)
• When heavily taxed, may show slightly more stutter than 16GB card in specific cases
• Given the small price gap, the 16GB provides a significantly better value proposition for most users

Nvidia GeForce RTX 5060

The RTX 5060 is Nvidia’s newest entry-level GPU for the RTX 50 series, and it brings a lot of high-end features to the budget segment. Built on the efficient Blackwell architecture and featuring 8GB of fast GDDR7 memory, this card is all about delivering excellent 1080p gaming performance while keeping power and heat under control. It uses the GB207 GPU with 3,840 CUDA cores and a modest 145W power draw — making it one of the most efficient cards in its class. Most models only require a single 8-pin connector, and the compact dual-fan designs from brands like ASUS, Palit, and MSI make it an easy fit for smaller builds or airflow-sensitive cases.

In terms of real-world gaming, the RTX 5060 consistently delivers 85–120 FPS at 1080p Ultra, depending on the game. Competitive titles like Valorant and CS2 easily exceed 200 FPS, while visually demanding games like Cyberpunk 2077 or Alan Wake 2 hold around 70–90 FPS at high settings with DLSS enabled. At 1440p, the 5060 performs well in most games, averaging 55–75 FPS depending on the title and whether DLSS is turned on. It's not a native 1440p powerhouse, but it can definitely hold its own — especially with DLSS 4 and Frame Generation switched on.

Speaking of DLSS, this card really shines when DLSS 4 is supported. DLSS Super Resolution combined with Frame Generation can effectively quadruple frame rates in compatible games. This means that in titles like Cyberpunk 2077, the RTX 5060 can output frame rates similar to what you’d expect from a much more expensive card — sometimes getting close to the RTX 4070’s performance when AI upscaling and framegen are active. That’s a huge win for budget-conscious gamers who want access to the latest Nvidia tech.

Compared to the RTX 5060 Ti, the base 5060 is about 12–17% slower on average at 1080p and 1440p. The Ti also has a 16GB variant, which may be worth considering if you want more VRAM for future titles. However, the standard 5060 delivers most of the same features — including ray tracing, DLSS 4, Reflex, and AV1 encoding — at a much lower cost and power draw. Versus the older RTX 4060, you’re looking at a 20–30% performance uplift, along with faster memory bandwidth and all the latest feature support.

The main drawback is, of course, the 8GB of VRAM. For now, it’s fine for 1080p gaming, and with DLSS it's possible to stretch it to 1440p. But if you’re planning to play modded games or ultra textures down the line, keep in mind that 8GB may become limiting — especially in future AAA titles. That said, if you’re building or upgrading a budget or mid-range system today, the RTX 5060 is an easy recommendation. It’s cool, quiet, power-efficient, and unlocks all of Nvidia’s latest visual enhancements at a price point that won’t hurt.

Take a look at a few of the RTX 5060 models we currently stock below:

View all RTX 5060 cards

Reasons to Buy

• Excellent 1080p performance in modern titles
• Very efficient (145W) – easy to cool and install in almost any build
• Access to all of Nvidia’s latest features without breaking the bank
• Quiet and compact

Reasons to Avoid

• 8GB VRAM might limit long-term use in future AAA titles at ultra settings
• Not ideal for native 1440p without DLSS – some settings compromises may be needed
• Only modest uplift from RTX 4060 in games that dont support dlss
• No 16GB option — need to jump to RTX 5060 Ti for more memory

Features and Trends to Consider

No discussion of 2025 GPUs is complete without touching on the major features and tech trends that define this generation. Beyond raw FPS numbers, you’ll want to consider these aspects when choosing a card:

DLSS 4 vs. FSR 4: Both Nvidia and AMD have fourth-gen upscaling tech now. DLSS 4 (Deep Learning Super Sampling 4.0) is Nvidia’s solution, available on RTX cards, which uses neural networks and dedicated tensor cores to upscale lower-resolution frames and also generate additional frames (Frame Generation). DLSS 4’s new Multi Frame Generation (MFG) can interpolate up to 3 extra frames for every 1 rendered, greatly boosting framerates. This can make a huge difference in demanding games – e.g., turning 30 FPS native into 90 FPS perceived. However, “fake” frames do add latency; Nvidia has worked to minimise it, but it’s something to be aware of (it’s usually fine for single-player, less ideal for competitive play). FSR 4 (FidelityFX Super Resolution 4.0) is AMD’s latest answer. Historically, FSR has been shader-based and didn’t require special hardware, but with FSR 4, AMD introduced AI-powered upscaling and frame generation as well. The catch: FSR 4’s most advanced mode is only supported on RX 9000-series GPUs (RDNA 4) – older cards can still use FSR 3.1 or FSR 2. Both DLSS4 and FSR4 aim to improve performance dramatically, but DLSS tends to have an edge in image quality at equivalent settings (thanks to more training data and optical flow calculations). That said, FSR 4 has reportedly closed the gap, offering image fidelity comparable to DLSS 2/3 in many cases. One advantage of FSR is broader support – e.g., FSR 3 (framegen) works on Nvidia and even older GPUs, whereas DLSS is proprietary to RTX cards. Bottom line: If you buy Nvidia, you get DLSS (widely supported, arguably the gold standard of upscaling). If you buy AMD, you rely on FSR (also widely supported and now improving). Both will make a mid-tier card perform like a higher-tier one in games that support them. It’s worth looking up if your favourite games use DLSS or FSR and factoring that in.

VRAM Requirements: We’ve touched on this throughout, but VRAM (video memory) has become a big talking point in recent years. Games in 2025, especially those with high-resolution textures and uncompressed assets, are hungry for VRAM. For 1080p gaming, 8GB has been the standard, but we are starting to see that 8GB can fall short if you max out everything (certain titles will use 9–10GB even at 1080p ultra). At 1440p, 8GB is even more likely to be a limit for max settings; 12GB is a safer minimum now, and 16GB is ideal to be safe. At 4K, 12GB can get used up quickly – 16GB or more is advisable for a high-end 4K card if you want ultra textures. This is precisely why Nvidia’s decision to offer a 16GB 5060 Ti was applauded. It’s also why AMD equipped even its midrange with 16GB. As a rule of thumb: more VRAM won’t directly increase FPS (unless you were running out), but running out of VRAM can crater your FPS (causing stutters or forcing lower texture quality). So, if you plan to keep your card for 3–4 years, it's safer to choose one with more VRAM to avoid limitations in future games. For example, an 8GB card might run today’s games fine, but 2-3 years down the line, you might regret not having 12–16GB when the latest titles come out. In summary: 16GB is the new 8GB in terms of comfort level, especially at 1440p. All the cards we discussed have at least 12GB except the 5060 Ti 8GB and the 5060. If you’re choosing between the 8GB and 16GB 5060 Ti, we strongly favour the 16GB for this reason.

Ray Tracing by Tier: Ray tracing has matured such that most new GPUs support it well, but performance varies by tier. The top-tier cards (5090, 5080) can often run ray-traced effects at 4K without needing to drop below 60 FPS (especially with DLSS). Mid-tier cards (5070 class) can usually handle ray tracing at 1440p, or 4K with upscaling. Entry-level cards (5060 Ti, 5060.) can do ray tracing at 1080p nicely, but at 1440p you’ll likely need upscaling or to dial back ray settings for smooth fps. Additionally, Nvidia still holds an advantage in ray tracing performance due to their hardware RT cores and many years of optimisation – for example, in a fully path-traced game like Cyberpunk 2077 RT Overdrive, an Nvidia RTX 50-series will significantly outpace an AMD card of equivalent tier. However, AMD narrowed the gap with RDNA 4, so much that at 1440p with moderate RT, the differences aren’t night and day. If playing a lot of ray-traced games is important to you, leaning Nvidia (or AMD’s highest tier like 9070 XT) is prudent. Also consider the level of RT: global illumination and path tracing are extremely heavy; simpler RT shadows or reflections are easier. Many mid-range cards can manage the latter, but struggle with the former without frame generation.

Efficiency and Thermals: As performance has gone up, power consumption unfortunately has too (especially at the high end). The RTX 5090’s 575W is testament to the brute-force approach for the flagship. Thankfully, not all cards are like that. Mid-range and budget GPUs this gen are quite power-efficient. Nvidia’s Ada and Blackwell architecture maintained good perf-per-watt, and AMD’s RDNA 4 also emphasised efficiency gains. What this means for you: ensure your power supply can handle the GPU you choose (for instance, 5090 needs ~1000W, 5080/5070 Ti class ~750W, 5070/9070 ~650W, 5060 Ti ~500W or less is fine). Also check the physical size and cooling solution – high-end cards often have 3-slot coolers that are 30+ cm long, which may not fit in smaller cases. The RTX 5090, while “compact” for what it is, is still a large card (about 3 slots thick). All that power also turns into heat. The 5090 and 5080 will dump a lot of heat into your case under load, so make sure you have good airflow. By contrast, something like a 5060 Ti or a 5060 can live in a mini-ITX case with a single fan and stay cool. Thermals are generally managed well by modern coolers – if you stick with trusted brands (ASUS, MSI, Gigabyte, etc.), their cooling solutions will keep the card within safe temps automatically by adjusting fan speeds. In summary, match your GPU choice to your case and PSU capabilities: an efficient mid-range card can make for a quiet, cool gaming rig, while a power-hungry beast will need a beefier setup (and will output a lot more heat into your room). If you’re not sure whether a card will fit inside your case, our team at CCL is always happy to help – just get in touch and we’ll check compatibility for you.

Other Features: Both vendors pack in other features that might sway you. Nvidia cards have things like ShadowPlay for easy gameplay recording, the best-in-class NVENC encoder (great if you stream or do video editing – though AMD’s encoder has gotten better), and support for technologies like Reflex (low-latency mode) and RTX IO (GPU-accelerated asset loading) which is similar to Microsoft’s DirectStorage. AMD cards offer features like Radeon Super Resolution (driver-level upscaling for any game), Anti-Lag+ (their version of latency reduction), and as mentioned, generally more VRAM for the money. Also, if you’re in the creative field, check if your preferred software benefits from CUDA (Nvidia) or if OpenCL support is fine (AMD). For most gamers, these ancillary features won’t overshadow core performance, but they’re good tie-breakers.

Tips for Choosing the Right GPU

By now we’ve covered the major players and features. But which one is right for you? Here are a few buying tips and things to consider when making your decision:

Match your GPU to your Monitor (Resolution & Refresh Rate): One of the simplest ways to choose is to consider your display. If you have a 1080p 60Hz monitor, you genuinely don’t need an RTX 5090 – a budget card like the 5060 will already max that out (and you might instead upgrade your monitor!). For a 1080p high refresh (120Hz/144Hz) setup, a mid-range card (5060 Ti, 5070) is ideal to push those high frame rates in esports titles. For 1440p 60Hz or 1440p 144Hz, look at mid-range to high-end: an RTX 5070 or RX 9070 will smoothly drive 1440p 60Hz in any game, and something like a 5070 Ti or 5080 is great for 1440p high refresh gaming. If you have a 4K 60Hz screen (like a 4K TV or 60Hz monitor), a mid-range card can suffice for 60 FPS at 4K with some settings tuning – but if you want Ultra settings all the time, an RTX 5080 or 5090 will give you more headroom. For those with a 4K high refresh (120Hz/144Hz) monitor, you’re in enthusiast territory – the RTX 5090 is really the only single GPU that can approach those frame rates in modern games (and even then, not 120 FPS in every game without DLSS). In summary: try to avoid over-buying or under-buying relative to your monitor. You’ll save money if you align the GPU’s strength with what you can actually see. CCL’s product listings often mention recommended resolutions for each GPU to help guide you.

Consider Future-Proofing vs. Budget: If you upgrade GPUs often (every 1-2 years), you can afford to be more lean and save money by buying just what you need right now. But if you plan to keep your card for many years, consider spending a bit more now to avoid being unhappy later. This often means choosing the model with more VRAM (e.g., 16GB vs 8GB) or a slightly higher tier to handle games 2-3 years from now. It’s a delicate balance – the GPU market evolves quickly, and there’s always something new around the corner, but certain things like memory capacity and architectural features (DLSS support, etc.) are forward-looking advantages you can secure now. Mid-range cards like the RTX 5070 are very future-proof for 1440p, whereas something like an 8GB card might show its age sooner. On the flip side, if your budget is tight, don’t feel pressured to stretch too far – even the “budget” cards today are extremely capable for mainstream gaming.

Brand and Model Selection: At CCL Computers, we stock GPUs from all the top add-in-board partners: ASUS, MSI, Gigabyte, Zotac, Palit, Inno3D, PNY, EVGA (though EVGA has exited GPUs recently), Sapphire, PowerColor, XFX, ASRock and more. Choosing a specific brand’s model can come down to cooler design, warranty, and personal preference. For example, ASUS ROG Strix cards are known for robust coolers (and typically higher price), while brands like Palit and Zotac often offer more compact models or value-oriented pricing. All the cards must adhere to baseline specs, so performance differences between models of the same GPU are usually small (a factory overclock might give ~2-5% extra fps). Focus on features: do you want RGB lighting? A smaller form factor? Extra HDMI ports? Check the model specifics. Also, consider the warranty and support – CCL as a retailer provides assistance with any RMA process, and brands like MSI, EVGA (if you find older stock), and Gigabyte often have 3-year or longer warranties. Buying from CCL also means you have UK-based support and you’re getting an officially sourced product (important for such an expensive item). We pride ourselves on customer service and can help with installation questions, troubleshooting, etc., which is something you don’t get if you buy off an auction site!

CCL Benefits: (Shameless plug incoming!) When you purchase from CCL Computers, you not only get competitive pricing but also peace of mind. We offer expert advice if you’re unsure which card fits your build – our team can help make compatibility checks (PSU, case, motherboard) and offer suggestions. We also frequently bundle games or offer promotions from Nvidia/AMD when available (for instance, you might get a free game code with certain GPUs – we’ll advertise this on the product page). Our shipping is fast and secure, so that shiny new GPU will reach you safely. And if anything goes wrong, our return and warranty support is hassle-free. Essentially, we aim to be more than just a seller; we want to be your partner in your PC building journey. So whether you’re grabbing a beastly RTX 5090 or a modest RTX 5060, you can trust CCL to provide a great buying experience.

Don’t Forget the Rest of Your System: A GPU is critical for gaming, but it needs a balanced system. Make sure your CPU is up to the task – a very weak or old CPU will bottleneck a high-end GPU, especially at low resolutions. For example, pairing a Ryzen 5 1400 with an RTX 5080 will hold back performance in many games. Generally, for mid/high GPUs, a modern 6-core or 8-core CPU (Intel 14th/14th-gen Core i5/i7 or AMD Ryzen 5/7 8000-series or 9000-series) is recommended to get the most out of the GPU. Also ensure you have enough RAM (16GB or 32GB) for gaming and a decent PSU as discussed. If you’re not sure, CCL offers pre-built systems and upgrade bundles – sometimes it might make sense to upgrade your platform alongside the GPU. We have a PC builder tool on our website to help select compatible components.

With these tips in mind, you should be well-equipped to choose the right graphics card. The summary table below compares the key stats of the GPUs we covered, which can also help inform your decision.

Comparison Table of GPUs (2025)

Conclusion

In 2025, gamers are spoilt with powerful GPU options at every tier. The Nvidia RTX 5090 is the uncontested king of performance – if you simply must have the highest FPS and money (and power draw) is no object, it’s the one to get for 4K and beyond. The RTX 5080 offers much of that high-end goodness at a relatively lower cost, making it a top pick for high-refresh 4K or 1440p gamers who want a card they won’t need to upgrade for a long time. In the mid-range, it’s a clash of titans: Nvidia’s RTX 5070 Ti and RTX 5070 versus AMD’s RX 9070 XT and RX 9070. These cards bring what used to be flagship-level performance down to prices that more gamers can afford. The 5070 Ti is great if you want 16GB and slightly better ray tracing, whereas the 9070 XT is great if you favour raw raster performance and extra VRAM – you can’t go wrong with either, as both will crush 1440p gaming. The RTX 5070 (12GB) and RX 9070 (16GB) are a close matchup too; Nvidia’s DLSS 4 magic versus AMD’s memory advantage – it’s a tie that comes down to personal preference and deals available. Lastly, for those on tighter budgets or gaming at 1080p, the RTX 5060 Ti (especially the 16GB variant) is a fantastic choice that ensures you can play the latest games smoothly without breaking the bank. And with the RTX 5060 now available, Nvidia has cemented its position in the budget segment — offering a more affordable way to access next-gen features and smooth 1080p gaming.

No matter which tier you go with, all these GPUs bring support for exciting new features like frame generation and advanced upscaling that make gaming smoother and more immersive. It’s also worth noting that the competition between Nvidia and AMD is intense – which is great news for gamers, as it drives innovation and sometimes more value. For example, AMD pushing larger VRAM forced Nvidia to respond in kind; Nvidia’s advancements in DLSS pushed AMD to develop FSR 4. We as gamers benefit from both ecosystems improving.

When making your final decision, consider the games you play, the resolution you use, and how long you want the card to last. Hopefully this guide has clarified the strengths and weaknesses of each option. If you’re still unsure, feel free to reach out to CCL’s team – we’re happy to provide personalised guidance. After all, we want you to be thrilled with your new graphics card, whether you’re chasing high scores in Counter-Strike 2, exploring detailed worlds in Cyberpunk 2077, or creating content with GPU-accelerated apps.

Here at CCL Computers, we stock all the latest GPUs discussed, and we pride ourselves on helping the community make the best PC hardware choices. We invite you to check out our catalogue, compare prices and specs, and even drop by our forums or social media to discuss with fellow enthusiasts.

What GPU are you most excited about in 2025? Do you have a favourite among the ones we listed, or maybe a dream build you’re aiming for? Let us know in the comments! We’d love to hear your thoughts or answer any questions you have. Happy gaming, and may your frames be high and your temperatures low!