

NVIDIA GeForce RTX 5090 wins 7 of 9 benchmarks, averaging 186.7% faster.
Both cards were measured first-party on our bench, same suite, same test rig.
The gap is widest in Stable Diffusion XL, where NVIDIA GeForce RTX 5090 leads by 343% (4.76 vs 21.08 images/min); the closest fight is Qwen3 4B (104% apart); VRAM decides part of this one: NVIDIA GeForce RTX 5090 runs 7 of our 12 AI workloads while the other card runs 3, models that don't fit score zero.
| Benchmark | NVIDIA GeForce RTX 3080 | NVIDIA GeForce RTX 5090 | Difference |
|---|---|---|---|
| Qwen3 4B tok/s | 184.08 | 375.6 | -51% |
| Llama 3.1 8B tok/s | 125.85 | 268.14 | -53% |
| Qwen2.5-Coder 14B tok/s | 0 | 149.75 | n/a |
| Qwen3 32B tok/s | 0 | 71.15 | n/a |
| Llama 3.3 70B tok/s | 0 | 0 | n/a |
| Stable Diffusion XL images/min | 4.76 | 21.08 | -77% |
| Z-Image Turbo images/min | 0 | 10.725 | n/a |
| FLUX.1 dev images/min | 0 | 2.036 | n/a |
| Qwen-Image-Edit images/min | 0 | 0 | n/a |
| Card | Per hour |
|---|---|
| NVIDIA GeForce RTX 3080 | $0.029 |
| NVIDIA GeForce RTX 5090 | $0.321 |
NVIDIA GeForce RTX 3080 is 11.07x cheaper per hour. Cheapest on-demand rate we see across RunPod and Vast.
| NVIDIA GeForce RTX 3080 | NVIDIA GeForce RTX 5090 | |
|---|---|---|
| VRAM | 10GB | 32GB |
| Transistors | 28,300M | 92,200M |
| Die size | 628.4 mm² | 750 mm² |
| Transistor density | 45 M/mm² | 122.9 M/mm² |
| Architecture | Ampere (GA102) | Blackwell (GB202) |
| Memory bandwidth | 760 GB/s | 1792 GB/s |
| Boost clock | 1,710 MHz | 2,407 MHz |
| TDP | 320 W | 575 W |
| Launch MSRP | $699 | $1,999 |
| Release | 2020-09-17 | 2025-01-30 |
NVIDIA GeForce RTX 3080 full review · NVIDIA GeForce RTX 5090 full review · All AI & Machine Learning rankings