Skip to main content
AI Choice Engine

GPU comparison

GeForce RTX 4090 vs GeForce RTX 5080

Compared on the specs that decide a purchase: memory, bandwidth, power, and which workloads each card actually fits — gaming, creative, or local models when that is the job.

Specifications verified October 2, 2026. Launch MSRP is not today's price — in September 2026 the memory shortage pushed street prices 1.2×–3× above it; the notes row below carries a dated lowest UK/US price where we have one.

GeForce RTX 4090

NVIDIA

GeForce RTX 4090

24GB VRAM · 1008 GB/s

vs
GeForce RTX 5080

NVIDIA

GeForce RTX 5080

16GB VRAM · 960 GB/s

Amazon UK search links are provided for current availability; no affiliate programme is currently configured, and no live price is claimed here. Affiliate disclosure.

GPU specification, memory, bandwidth, and price comparison
SpecificationGeForce RTX 4090GeForce RTX 5080
VendorNVIDIANVIDIA
MemoryWinner: 24GB16GB
Usable for a modelWinner: 24GB16GB
BandwidthWinner: 1008 GB/s960 GB/s
Segmentconsumerconsumer
ReleasedOctober 12, 2022January 30, 2025
Launch MSRP~$1,599~£1,184~$999~£740
Notes24GB GDDR6X and 1008GB/s theoretical bandwidth. NVIDIA's launch notice lists October 12, 2022 availability starting at $1,599. Current production status, retailer stock and used prices are unverified here. Local-model fit is a planning estimate; model files, context, cache, backend and driver support determine actual memory use and speed.16GB GDDR7; NVIDIA specifies 960GB/s theoretical bandwidth. Launch starting price $999 and announced availability January 30, 2025. Local-model fit is a planning estimate; model files, context, cache, backend and driver support determine actual memory use and speed.
UK street priceSearch Amazon (opens in new tab)Search Amazon (opens in new tab)

Local models

Estimated local-model fit

Planning estimates from published VRAM and a weights-plus-overhead heuristic. Context length, KV cache, quantization and runtime can change actual memory needs.

Local model fit comparison for both GPUs
ModelGeForce RTX 4090 (est.)GeForce RTX 5080 (est.)
Qwen 3 4B4B parametersEst. fitEst. fit
Qwen 3.8 27B27B parametersEst. fitEst. no fit
Gemma 4 31B30.7B parametersEst. tightEst. no fit
Llama 3.1 8B Instruct8B parametersEst. fitEst. fit
Qwen 3 14B14.8B parametersEst. fitEst. fit
Gemma 3 27B27B parametersEst. fitEst. no fit
Qwen 3 30B-A3B30.5B parametersEst. tightEst. no fit
Qwen 3 32B32.8B parametersEst. tightEst. no fit
Llama 3.3 70B70B parametersEst. no fitEst. no fit
Muse Glimmer 30B29.6B parametersEst. tightEst. no fit
Mixtral 8x22B141B parametersEst. no fitEst. no fit
Qwen 3 235B-A22B235B parametersEst. no fitEst. no fit
Qwen 3 Coder 480B-A35B480B parametersEst. no fitEst. no fit
DeepSeek V4-Flash284B parametersEst. no fitEst. no fit

Common questions

GeForce RTX 4090 vs GeForce RTX 5080

Answered from the verified figures on this page rather than general guidance.

Which costs less at launch MSRP, GeForce RTX 4090 or GeForce RTX 5080?
GeForce RTX 5080 launched lower at ~$999 against ~$1599 for GeForce RTX 4090. Launch MSRP is an anchor, not a live price.
Which is better for high-refresh gaming, GeForce RTX 4090 or GeForce RTX 5080?
GeForce RTX 4090 has more memory bandwidth — 1008 GB/s against 960 GB/s. That usually helps high-refresh 1440p and 4K once VRAM is sufficient for your settings. Still verify with dated independent tests for the games you play.
Which has more VRAM for gaming and creative work, GeForce RTX 4090 or GeForce RTX 5080?
GeForce RTX 4090 — 24GB usable against 16GB on GeForce RTX 5080. Extra memory helps texture mods, video timelines, and 3D scenes before it helps esports frame rates.
For local inference only — which holds larger models, GeForce RTX 4090 or GeForce RTX 5080?
GeForce RTX 4090. With 24GB usable against 16GB it additionally holds Qwen 3.8 27B, Gemma 4 31B, Gemma 3 27B, Qwen 3 30B-A3B, Qwen 3 32B, Muse Glimmer 30B at Q4.
For local inference only — is GeForce RTX 4090 or GeForce RTX 5080 faster at generating tokens?
GeForce RTX 4090, at 1008 GB/s against 960 GB/s — roughly 1.1× the memory bandwidth. Bandwidth is the practical ceiling on token throughput once a model fits, but it only matters if the model fits in the first place.
For local inference only — can either run a 70B model?
Neither. A 70B model needs about 48GB at Q4 including overhead, above both cards. The largest that fits is Qwen 3 32B on GeForce RTX 4090 and Qwen 3 14B on GeForce RTX 5080.

If one card holds a model the other cannot, that capacity difference usually outweighs a bandwidth advantage. A model that spills to system memory can be substantially slower, but the actual impact depends on the backend, transfer path, context, and workload.

All GPU comparisonsCompare hosted models instead