DreamShaper XL Turbo · 10 GPUs measured first-party · text-to-image · Updated October 2026
What GPU Do You Need for DreamShaper XL Turbo?
DreamShaper XL Turbo on 10 GPUs, measured first-party: NVIDIA GeForce RTX 5090 leads at 70.72 images/min, RTX 4060 Ti 16GB trails at 16.4 images/min, and it peaked at 13GB of VRAM.
Benchmarked weights: Lykon/dreamshaper-xl-v2-turbo
NVIDIA GeForce RTX 5090
70.72 images/min on DreamShaper XL Turbo, the ceiling. Measured on our bench. 32GB of VRAM, $1,999 at launch.
- 70.72 images/min on DreamShaper XL Turbo
- 32GB, clears the DreamShaper XL Turbo floor
- 575W board rating
NVIDIA GeForce RTX 4090
54.35 images/min on DreamShaper XL Turbo, fastest card you can buy at retail. Measured on our bench. 24GB of VRAM, $1,599 at launch.
- 54.35 images/min on DreamShaper XL Turbo
- 24GB, clears the DreamShaper XL Turbo floor
- 450W board rating
NVIDIA GeForce RTX 3060
10.14 images/min on DreamShaper XL Turbo, lowest launch price that still fits. Measured on our bench. 12GB of VRAM, $329 at launch.
- 10.14 images/min on DreamShaper XL Turbo
- 12GB, clears the DreamShaper XL Turbo floor
- 170W board rating
NVIDIA GeForce RTX 4070 Super
26.34 images/min on DreamShaper XL Turbo, most speed per dollar. Measured on our bench. 12GB of VRAM, $599 at launch. That is 43.97 images/min per $1,000 of launch price.
- 26.34 images/min on DreamShaper XL Turbo
- 12GB, clears the DreamShaper XL Turbo floor
- 220W board rating
What GPU Do You Need for DreamShaper XL Turbo?, images/min by GPU
Efficiency: images/min per 100W drawn
Power is the average pulled during the run, sampled at 1Hz. The fastest card is often not the one here, and for anything left running this is the number that shows up on the bill.
Value: images/min per $1,000 of MSRP
Launch price, not street price, so it ages. A speed leaderboard always crowns the most expensive card; this is the counterweight.
DreamShaper XL Turbo. Measured image generation speed by GPU
| GPU | Images/min | s per image | img/W·min | Avg power |
|---|---|---|---|---|
| NVIDIA GeForce RTX 5090 | 70.72 | 0.848 | 0.132 | 534.8 W |
| NVIDIA GeForce RTX 4090 | 54.35 | 1.104 | 0.132 | 411.7 W |
| NVIDIA GeForce RTX 4070 Ti Super | 30.49 | 1.968 | 0.141 | 216.6 W |
| NVIDIA GeForce RTX 3080 Ti | 28.09 | 2.136 | 0.072 | 391.9 W |
| NVIDIA GeForce RTX 3090 | 27.3 | 2.198 | 0.079 | 343.4 W |
| NVIDIA GeForce RTX 4070 Super | 26.34 | 2.278 | 0.132 | 199.6 W |
| NVIDIA GeForce RTX 4070 | 22.67 | 2.647 | 0.116 | 194.6 W |
| GeForce RTX 5060 Ti | 18.31 | 3.277 | 0.118 | 154.7 W |
| NVIDIA GeForce RTX 4060 Ti 16GB | 16.44 | 3.65 | 0.112 | 146.6 W |
| NVIDIA GeForce RTX 3060 | 10.14 | 5.919 | 0.066 | 154.4 W |
What the numbers show. Across 4 GPUs measured on our own bench, RTX 5090 is fastest at 70.7 images/min. The slowest, RTX 4060 Ti 16GB, manages 16.4, so the spread is 4.3x from top to bottom. The fastest card with 16GB or less is RTX 4060 Ti 16GB at 16.4 images/min.
About DreamShaper XL Turbo. DreamShaper XL Turbo: from Lykon, 2.6B parameters, on Hugging Face since February 2024, OpenRAIL++ licence. 52,394 downloads in the last 30 days.
How it compares. RTX 4090: DreamShaper XL Turbo 54.35 images/min, Playground v2.5 9.96 (3B), SDXL Turbo 652.1 (3B), Stable Diffusion 3 Medium 13.95 (2B), FLUX.2 klein 4B 42.38 (4B). 1 of 4 beat DreamShaper XL Turbo here.
Cost on a rented GPU. 1,000 images of DreamShaper XL Turbo: $0.059 on a RTX 3060 ($0.036/hr, 99 min), $0.092 on a RTX 5090 ($0.39/hr, 14 min, 1.5x the cost).
DreamShaper XL Turbo: cost per 1,000 images on rented GPUs
| GPU | Cheapest rate | Speed (images/min) | Cost per 1,000 images |
|---|---|---|---|
| NVIDIA GeForce RTX 3060 | $0.036/hr | 10.14 | $0.059 |
| NVIDIA GeForce RTX 3090 | $0.12/hr | 27.3 | $0.074 |
| NVIDIA GeForce RTX 4070 Ti Super | $0.16/hr | 30.49 | $0.089 |
| NVIDIA GeForce RTX 4070 | $0.12/hr | 22.67 | $0.090 |
| NVIDIA GeForce RTX 5090 | $0.39/hr | 70.72 | $0.092 |
| NVIDIA GeForce RTX 4090 | $0.34/hr | 54.35 | $0.10 |
| NVIDIA GeForce RTX 3080 Ti | $0.18/hr | 28.09 | $0.11 |
| GeForce RTX 5060 Ti | $0.14/hr | 18.31 | $0.12 |
Cheapest hourly rate we track on RunPod and Vast.ai, divided by the measured speed. Startup time and storage are extra.
Speed tiers for DreamShaper XL Turbo. 30+ images/min: 3 (RTX 5090, RTX 4090, RTX 4070 Ti Super); 6-30 images/min: 7 (RTX 3080 Ti, RTX 3090, RTX 4070 Super). 30 images/min means two seconds or less per picture.
Time per image. DreamShaper XL Turbo at 1024px and 8 steps: 0.8s per image on the RTX 5090, 5.9s on the RTX 3060. A batch of 100 takes 1 min on the fastest card and 10 min on the slowest.
VRAM for DreamShaper XL Turbo. Measured peak 10.6GB, so 12GB is the smallest common card size; smallest card it ran on: RTX 3080 Ti (12GB).
Power on DreamShaper XL Turbo. Most efficient: RTX 4070 Ti Super, 217W, 0.12 kWh per 1,000 images. Hungriest: RTX 5090, 535W, 0.13 kWh. At $0.15/kWh: $0.018 per 1,000 images.
Our verdict
Fastest on DreamShaper XL Turbo: NVIDIA GeForce RTX 5090, 70.72 images/min. Cheapest consumer card that ran it: NVIDIA GeForce RTX 3060 ($329, 10.14 images/min). Cheapest to rent per job: NVIDIA GeForce RTX 3060, $0.059 per 1,000 images.
FAQ
- What GPU do I need to run DreamShaper XL Turbo?
- About 11GB. Cheapest consumer card that ran it: NVIDIA GeForce RTX 3060 (12GB, 10.14 images/min).
- How fast is DreamShaper XL Turbo on the NVIDIA GeForce RTX 5090?
- 70.72 images/min at 535W, faster than every datacenter card we ran it on.
- How much does it cost to run DreamShaper XL Turbo in the cloud?
- $0.059 per 1,000 images on a NVIDIA GeForce RTX 3060 at $0.036/hr, cheapest of 8 rentable cards we measured.
- Can I run DreamShaper XL Turbo on a 12GB, 16GB or 24GB card?
- It used 10.6GB at the precision we tested. 12GB: yes; 16GB: yes; 24GB: yes.
- Is the RTX 4090 or the RTX 3090 faster for DreamShaper XL Turbo?
- The RTX 4090: 54.35 vs 27.3 images/min, 99% faster on our bench.
- Should I buy or rent a RTX 5090 for DreamShaper XL Turbo?
- Its $1,999 launch price buys 5,139 rented hours at $0.39/hr, enough for about 21,805x 1,000 images of DreamShaper XL Turbo. Buy only if you'll run more than that.
How we test
DreamShaper XL Turbo at 1024px in diffusers, bf16, native precision with no offload, timed over three generations after a warmup, with power and VRAM sampled throughout. Image speed tracks tensor compute and architecture generation more than memory bandwidth, so the order here differs from our LLM boards.