RTX 4090 vs RTX 5080
The complete head-to-head: full specs, real per-game FPS with 1% lows, ray tracing, a live CPU-bottleneck check, thermals, power and value, with a straight verdict.
GeForce RTX 4090
GeForce RTX 5080
avg, 1080p
The short answer
This is the honest answer to a question a lot of shoppers ask: does the new x80 beat the old flagship? Not quite. The RTX 4090 is still about 12% faster on average than the RTX 5080, and it carries 24 GB of VRAM against the 5080's 16 GB. What the 5080 offers instead is a much lower price, roughly 37% less at street value, far less power draw, and a newer feature set including DLSS 4 Multi Frame Generation. If raw performance and maximum VRAM matter most, the 4090 still wins outright. If value, power, and a smaller footprint matter more, the 5080 is the smarter buy.
The RTX 4090 was the undisputed flagship of the last generation, and the RTX 5080 is this generation's second-tier card, one rung below the RTX 5090. Comparing them answers a real question: is it worth chasing down a discontinued flagship, or is the new generation's x80 tier good enough. The honest answer is that the 4090 still wins on raw speed and memory, but the 5080 wins clearly on price, power, and modern features.
This page lines the two up head to head across ten games plus synthetic and creator benchmarks, breaks down what DLSS 3 Frame Generation versus DLSS 4 Multi Frame Generation actually changes on screen, runs a live CPU-pairing check so you know whether a weaker processor would hold either card back, and tracks price history so you can see whether the 4090's premium over the 5080 is closing or holding steady.
At a glance
Seven dimensions, two cards. Bigger area is not automatically better, each card owns different corners.
Strengths radar
Each axis scored 0-100 relative to the pair
Performance vs price
Up and to the left is better value
Grey dots are neighbouring cards for context.
Which should you buy?
Pick your use case, the winner changes with what you actually do.
Still the faster card overall, and the gap holds or widens slightly as resolution climbs toward 4K.
Newer 4th-gen RT cores and DLSS 4's transformer model let the 5080 nearly close the raster gap in the heaviest path-traced titles.
24 GB against 16 GB is a real gap for local AI. The 4090 fits noticeably larger models and batch sizes.
More VRAM headroom for large timelines and a small edge from the extra core power, though both encode AV1 well.
About 37% cheaper at street price and drawing 90 W less, the 5080 delivers most of the flagship experience for a lot less money and heat.
More bandwidth and a wider bus give the 4090 the edge in the heaviest, highest-internal-resolution emulation workloads.
Already own an RTX 4090? Don't trade down for a 5080.
Short answer: probably not
The RTX 5080 is a newer, cheaper, more efficient card, but it is not a faster one. Trading a 4090 for a 5080 means giving up 12% average performance and 8 GB of VRAM to save money you have already spent. That only makes sense if you specifically need to cut power draw or want to sell the 4090 while its resale value is still strong.
If you want a genuine step up from either card, the RTX 5090 is the only option, and it costs meaningfully more than both.
Full specification comparison
Every spec side by side. The highlighted figure wins each row.
| Specification | RTX 4090 | RTX 5080 |
|---|---|---|
| Architecture | ||
| GPU / architecture | AD102 · Ada Lovelace | GB203 · Blackwell |
| Process node | TSMC 4N (5 nm class) | TSMC 4NP (5 nm class) |
| Release date | Oct 12, 2022 | Jan 30, 2025 |
| Launch MSRP | $1,599 | $999 ▲ |
| Compute | ||
| CUDA cores | 16,384 ▲ | 10,752 |
| RT cores | 128 · 3rd gen ▲ | 84 · 4th gen |
| Tensor cores | 512 · 4th gen ▲ | 336 · 5th gen |
| Base / boost clock | 2235 / 2520 MHz | 2295 / 2617 MHz ▲ |
| L2 cache | 72 MB ▲ | 64 MB |
| Memory | ||
| VRAM | 24 GB GDDR6X ▲ | 16 GB GDDR7 |
| Memory bus | 384-bit ▲ | 256-bit |
| Bandwidth | 1,008 GB/s ▲ | 960 GB/s |
| Power & form factor | ||
| TDP | 450 W | 360 W ▲ |
| Power connector | 1× 16-pin (12VHPWR) | 1× 16-pin (12V-2x6) |
| Suggested PSU | 850 W | 850 W |
| PCIe interface | 4.0 ×16 | 5.0 ×16 ▲ |
| Features | ||
| Upscaling | DLSS 3 + Frame Gen | DLSS 4 + Multi Frame Gen ▲ |
| AV1 encode (NVENC) | Yes · 2 encoders | Yes · 2 encoders |
| Display outputs | 3×DP1.4a · HDMI 2.1 | 3×DP2.1 · HDMI 2.1b ▲ |
Average FPS by game
Switch resolution, metric, ray tracing and upscaling, the bars update live. Hover a bar for the avg / 1% / 0.1% breakdown.
Average FPS · 1080p · Rasterization. Representative data, hover a bar for avg / 1% / 0.1%. Note the RTX 5080's 1% lows fall off at 1440p and up from its 8 GB VRAM.
Smoothness & frametimes
Average FPS hides stutter. This is one run plotted as frametime, a flat line is smooth, spikes are hitches. Watch what the RTX 5080's 8 GB VRAM does in a texture-heavy game.
Frametime: Alan Wake 2 with path tracing, 4K
Lower and flatter is better · 16.7 ms = 60 fps
Counterintuitive but real: the slower RTX 4090 delivers a smoother line here because 12 GB holds the texture pool the 8 GB RTX 5080 keeps evicting, those spikes are VRAM stutter, not raw speed.
Lab scores
Clean apples-to-apples tests, useful, but weaker at predicting real games than the FPS above.
Ray tracing & upscaling
Ray tracing tanks both cards at native resolution. What makes it playable is upscaling, and here the RTX 5080 has a real edge.
Is frame generation "real" performance?
The RTX 4090 supports DLSS 3's single Frame Generation, one AI frame inserted between two real ones. The RTX 5080 runs DLSS 4's Multi Frame Generation, which can insert up to three AI frames per real one, a meaningful gap on a high-refresh monitor even though the 5080 trails in raw raster fps.
Frame generation works best layered on top of an already healthy base frame rate, 50-60 fps or more, and both cards clear that bar easily in most games. Turn it off for competitive shooters on either card, the added latency works against you there regardless of which version you are running.
Will your CPU bottleneck these cards?
A faster GPU only helps if your CPU keeps up. Pick your CPU and resolution to see expected FPS and bottleneck for each card.
CPU + GPU pairing check
The willitbottleneck engine, built right into the comparison.
Estimated from a 43-CPU model across 1440p. A weak CPU mostly bites at 1080p; at higher resolutions the GPU becomes the limiter and the bottleneck shrinks.
Thermals & noise
The efficiency gap shows up here: the newer card runs cooler and quieter, day in and day out.
Real draw, efficiency & PSU
TDP is the rated figure; what matters is real gaming draw and how many frames you get per watt.
Both use a single 8-pin connector. If your PSU already ran the RTX 4090, it will run the RTX 5080 easily. Leave about 30% headroom for spikes.
Will it fit, and what to pair it with?
Size to scale
Card length matters for small cases
What monitor to pair
RTX 4090: excellent at 4K 100-144 Hz and the best card for 1440p at 240 Hz-plus. RTX 5080: strong at 4K 60-100 Hz and very comfortable at 1440p 165 Hz-plus. Both are capable 4K cards; the 4090 simply has more headroom left over at the top end.
Running an older PCIe 3.0 board?
The 5080 runs PCIe 5.0 ×16, but no current game or motherboard chipset saturates that bandwidth, so it is future headroom rather than a real gain today. The 4090's PCIe 4.0 ×16 costs nothing in practice on any modern board.
Price history & dollars-per-frame
Raw price is only half the story. Cost per frame is the number that decides it, and it keeps moving as prices fall.
New price over time
Street price of new cards (USD)
The 24 GB vs 16 GB question
RTX 4090 · 24 GB
- 24 GB comfortably covers every game and most professional workloads without hitting a ceiling.
- Local AI and large 3D scenes benefit directly, 24 GB fits noticeably bigger models and batches than the 5080's 16 GB.
- 384-bit bus and 1,008 GB/s bandwidth stay ahead of the 5080 despite being a generation older.
- The premium for that extra memory is steep, the 4090 costs well over 50% more than the 5080 at current street prices.
RTX 5080 · 16 GB
- 16 GB is still plenty for 1440p and most 4K gaming today, and GDDR7 gives it 960 GB/s despite a narrower 256-bit bus.
- Newer architecture and DLSS 4 Multi Frame Generation claw back real-world playability even where raw fps trails the 4090.
- 16 GB is the first ceiling you will hit in the newest AAA games at 4K with texture packs and ray tracing maxed.
- For local AI and the largest creative workloads, 16 GB is a real limitation next to the 4090's 24 GB.
Emulation & memory bandwidth
The narrower bus is a real downgrade for emulation
The RTX 4090's 384-bit bus and 1,008 GB/s of bandwidth give it a real edge in the heaviest RPCS3 and high-internal-resolution Switch emulation over the 5080's 960 GB/s. Both are far more than enough for the vast majority of emulation use cases, the gap only shows up at extreme upscale factors most players never push to.
Creator, AI & streaming
The newer card is quicker per task, but the older card's extra VRAM unlocks larger jobs it can't fit at all.
Streaming: the AV1 encoder is the differentiator
Both cards have dual hardware AV1 encoders, so stream-and-record-at-once workflows run equally well on either. The 4090's larger VRAM pool gives it more headroom on huge multi-layer timelines, while the 5080's newer encoder block trades blows with it on export speed for shorter, simpler projects.
Overclocking & undervolting
Both cards ship close to their thermal and power limits from the factory, leaving modest headroom, roughly 7-8%, for a traditional overclock. Undervolting is the better use of tuning time on either: similar performance with a real drop in temperature, noise, and total system power draw.
Pros & cons
RTX 4090
- About 12% faster on average, more at 4K
- 24 GB of VRAM leads every AI and creator workload here
- More bandwidth for the heaviest emulation and rendering tasks
- Costs roughly 55-60% more than the 5080, new or used
- Draws 90 W more, needs a bigger PSU and cooling budget
RTX 5080
- Around 37% cheaper at street price, a large gap
- 360 W is far easier on PSU headroom and case airflow
- DLSS 4 Multi Frame Generation and newer RT cores close the gap in ray tracing
- About 12% slower on average, more at 4K
- 16 GB of VRAM is the first thing you will outgrow versus the 4090's 24 GB
Buying pitfalls
Don't assume newer beats older at this tier
It is easy to assume a new-generation card automatically beats last generation's flagship. Here it does not, the discontinued RTX 4090 is still about 12% faster than the current RTX 5080. Buy on your actual priorities, price and power versus raw speed and VRAM, not on release date.
4090 supply keeps shrinking
Production stopped when the 50-series launched, so remaining new 4090 stock is limited and used listings vary widely in condition and price. Buy from a seller with a return window, and confirm the card has not been used for sustained mining or AI compute at aggressive power limits.
RTX 4090 vs RTX 5080: your questions answered
Not on raw speed. The discontinued RTX 4090 still runs about 12% faster on average and keeps 24 GB of VRAM against the 5080's 16 GB. What the 5080 wins outright is the price tag, roughly 37% less at street value, plus a 90 W lighter power draw and DLSS 4 Multi Frame Generation.
Grab the 4090 when top-end performance or the extra VRAM headroom for AI and creator work is the deciding factor and a fairly priced unit turns up. Grab the 5080 when you'd rather spend less, run cooler, and still get most of the flagship experience out of a current-generation card.
Absolutely. It outruns the current RTX 5080 in raster and chews through native 4K ultra without breaking a sweat. Being out of production doesn't make it obsolete, it just means the remaining supply is tighter.
Generally no. That swap costs you 12% average performance and 8 GB of VRAM in exchange for cash you've already spent. It only starts to make sense if cutting 90 W of power draw or cashing out the 4090's still-strong resale value matters more to you than raw speed.
The 4090 takes this one without much argument. 24 GB of VRAM handles noticeably larger local models and batch sizes than the 5080's 16 GB allows, and its extra memory bandwidth keeps those memory-hungry workloads fed.
Nearly a wash. Both pack dual AV1 hardware encoders, but the 4090's bigger VRAM pool gives it room to breathe on massive multi-layer timelines, while the 5080's newer encoder holds its own on smaller, shorter export jobs.
Yes, each has dedicated RT hardware. The 5080's cores are a generation newer, and paired with DLSS 4 they let it close in on or even edge past the 4090 in the most demanding path-traced games, even though the 4090 leads in plain rasterization.
The 5080 runs DLSS 4's Multi Frame Generation, which can add up to three AI-generated frames for every real one rendered. The 4090 tops out at DLSS 3's single Frame Generation, a smaller boost but still a useful one on a high-refresh display.
It's possible at 1080p and sometimes even 1440p if the processor is dated. Run the pairing checker above; something in the Ryzen 7 7800X3D or Core i9-14900K class keeps up with either GPU, and matters even more when the CPU is paired with the faster 4090.
The 4090 pulls around 430 W under real gaming loads versus roughly 330 W for the 5080, and that gap shows up directly in temperatures, the 5080 runs noticeably cooler and quieter. Neither spins its fans at idle.
NVIDIA lists 850 W as the recommended supply for both the 4090 and the 5080, each fed through a single 16-pin connector. An ATX 3.0-rated unit in the 850 to 1000 W range gives comfortable margin on either, and covers the 4090's larger transient power spikes especially well.
Check your clearance either way, the 4090 runs about 304 mm across 3 slots and most 5080 cards land around 300 mm and 3 slots, close enough that neither is the easy fit. The 5080's PCIe 5.0 support is forward-looking only; running the 4090 on a PCIe 4.0 board costs nothing in practice.
Not on equal footing. The mobile chip badged RTX 4090 is a much smaller part than the desktop card it shares a name with, while a laptop RTX 5080 does exist and, like every mobile GPU, falls well short of its desktop counterpart. Only compare laptop to laptop.
Both are genuine 4K cards. The 4090 has more room to spare for native 4K ultra with ray tracing cranked all the way, and the 5080 reaches similar results at the same settings by leaning a bit more on DLSS.
A used 4090 near $1,400 pays off mainly if you specifically need that extra VRAM or the top-tier speed. Most people come out ahead with a new 5080 around $979, full warranty included, lower power draw, and current-gen features for meaningfully less cash.
No reason to expect that. Both sit on architectures current enough to keep getting full Game Ready driver updates for years to come, so driver longevity isn't a factor in choosing between this pair.