Can My GPU Run Seedream 3.0?
Text to Image · VRAM requirements and compatibility for 78 GPUs
Home › AI Models › Seedream 3.0
ByteDance internal image generation model. Seedream 3.0 is a text to image model — developed by ByteDance, released 2025-03, 8B parameters, 16.0 GB download, license: research.
Short answer: you need at least 8 GB of VRAM (12 GB recommended). 76 of 78 GPUs can run it; the smallest NVIDIA card that can is the GTX 1070 (8 GB).
Check Seedream 3.0 on your GPUSeedream 3.0 VRAM requirements
| Format | VRAM |
|---|---|
| FP16 (full precision) | 16 GB |
| FP8 | 10 GB |
| Minimum | 8 GB |
| Recommended | 12 GB |
| Optimal | 24 GB |
Tip: ByteDance T2I model. Weights NOT publicly available, API only (Together AI, etc).
Available as: Diffusers, FP8 versions, CPU offloading.
Official pages: GitHub
Which GPUs can run Seedream 3.0?
Ratings go from S (perfect) to F (can't run) and use the best format that fits in each card's VRAM. How ratings work.
NVIDIA
| GPU | VRAM | Rating | Runs as | Speed | |
|---|---|---|---|---|---|
| RTX 5090 | 32 GB | Perfect | FP16 | Fast | Details |
| RTX 5080 | 16 GB | Good | FP16 | Fast | Details |
| RTX 5070 Ti | 16 GB | Good | FP16 | Fast | Details |
| RTX 5070 | 12 GB | Good | FP8 | Moderate | Details |
| RTX 5060 Ti 16GB | 16 GB | Usable | FP16 | Moderate | Details |
| RTX 5060 Ti 8GB | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 5060 | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 4090 | 24 GB | Great | FP16 | Fast | Details |
| RTX 4080 Super | 16 GB | Good | FP16 | Fast | Details |
| RTX 4080 | 16 GB | Good | FP16 | Fast | Details |
| RTX 4070 Ti Super | 16 GB | Good | FP16 | Fast | Details |
| RTX 4070 Ti | 12 GB | Good | FP8 | Moderate | Details |
| RTX 4070 Super | 12 GB | Good | FP8 | Moderate | Details |
| RTX 4070 | 12 GB | Good | FP8 | Moderate | Details |
| RTX 4060 Ti 16GB | 16 GB | Usable | FP16 | Moderate | Details |
| RTX 4060 Ti 8GB | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 4060 | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 3090 Ti | 24 GB | Great | FP16 | Fast | Details |
| RTX 3090 | 24 GB | Great | FP16 | Fast | Details |
| RTX 3080 Ti | 12 GB | Good | Offload | Very Slow | Details |
| RTX 3080 12GB | 12 GB | Good | Offload | Very Slow | Details |
| RTX 3080 10GB | 10 GB | Usable | Offload | Very Slow | Details |
| RTX 3070 Ti | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 3070 | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 3060 Ti | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 5050 | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 3060 | 12 GB | Usable | Offload | Very Slow | Details |
| RTX 2080 Ti | 11 GB | Usable | Offload | Very Slow | Details |
| RTX 2080 Super | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 2080 | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 2070 Super | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 2070 | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 2060 Super | 8 GB | Barely | Offload | Very Slow | Details |
| RTX 2060 | 6 GB | Can't Run | — | — | Details |
| GTX 1080 Ti | 11 GB | Usable | Offload | Very Slow | Details |
| GTX 1080 | 8 GB | Barely | Offload | Very Slow | Details |
| GTX 1070 Ti | 8 GB | Barely | Offload | Very Slow | Details |
| GTX 1070 | 8 GB | Barely | Offload | Very Slow | Details |
| GTX 1060 6GB | 6 GB | Can't Run | — | — | Details |
AMD
| GPU | VRAM | Rating | Runs as | Speed | |
|---|---|---|---|---|---|
| RX 9070 XT | 16 GB | Usable | FP16 | Fast | Details |
| RX 9070 | 16 GB | Barely | FP16 | Moderate | Details |
| RX 7900 XTX | 24 GB | Usable | FP16 | Fast | Details |
| RX 9070 GRE | 12 GB | Usable | FP8 | Moderate | Details |
| RX 7900 XT | 20 GB | Usable | FP16 | Fast | Details |
| RX 9060 XT 16GB | 16 GB | Barely | FP16 | Moderate | Details |
| RX 9060 XT 8GB | 8 GB | Can't Run | Offload | Very Slow | Details |
| RX 7900 GRE | 16 GB | Barely | FP16 | Moderate | Details |
| RX 7800 XT | 16 GB | Barely | FP16 | Moderate | Details |
| RX 7700 XT | 12 GB | Barely | Offload | Very Slow | Details |
| RX 7600 XT | 16 GB | Barely | FP16 | Moderate | Details |
| RX 7600 | 8 GB | Can't Run | Offload | Very Slow | Details |
| RX 6950 XT | 16 GB | Barely | FP16 | Moderate | Details |
| RX 6900 XT | 16 GB | Barely | FP16 | Moderate | Details |
| RX 6800 XT | 16 GB | Barely | FP16 | Moderate | Details |
| RX 6800 | 16 GB | Barely | FP16 | Moderate | Details |
| RX 6700 XT | 12 GB | Barely | Offload | Very Slow | Details |
| RX 6600 XT | 8 GB | Can't Run | Offload | Very Slow | Details |
| RX 6600 | 8 GB | Can't Run | Offload | Very Slow | Details |
Apple
| GPU | VRAM | Rating | Runs as | Speed | |
|---|---|---|---|---|---|
| M5 Ultra | 96 GB | Usable | FP16 | Fast | Details |
| M3 Ultra | 96 GB | Usable | FP16 | Fast | Details |
| M4 Max | 48 GB | Usable | FP16 | Fast | Details |
| M5 Max | 48 GB | Usable | FP16 | Fast | Details |
| M4 Pro | 24 GB | Barely | FP16 | Moderate | Details |
| M5 Pro | 24 GB | Barely | FP16 | Moderate | Details |
| M4 | 16 GB | Can't Run | FP16 | Slow | Details |
| M3 Max | 36 GB | Usable | FP16 | Moderate | Details |
| M3 Pro | 18 GB | Barely | FP16 | Slow | Details |
| M5 | 16 GB | Can't Run | FP16 | Slow | Details |
| M6 | 16 GB | Can't Run | FP16 | Slow | Details |
| M3 | 16 GB | Can't Run | FP16 | Slow | Details |
| M2 Ultra | 192 GB | Usable | FP16 | Fast | Details |
| M2 Max | 96 GB | Usable | FP16 | Fast | Details |
| M2 Pro | 32 GB | Barely | FP16 | Moderate | Details |
| M2 | 24 GB | Barely | FP16 | Slow | Details |
| M1 Ultra | 128 GB | Usable | FP16 | Fast | Details |
| M1 Max | 64 GB | Usable | FP16 | Moderate | Details |
| M1 Pro | 32 GB | Barely | FP16 | Moderate | Details |
| M1 | 16 GB | Can't Run | FP16 | Slow | Details |
Frequently asked questions
How much VRAM does Seedream 3.0 need?
Seedream 3.0 needs at least 8 GB of VRAM, 12 GB recommended, and 16 GB to run at full FP16 precision.
Can Seedream 3.0 run on 8 GB of VRAM (e.g. NVIDIA RTX 4060)?
On the NVIDIA RTX 4060: yes, with CPU offloading (rating D, barely).
Can Seedream 3.0 run on the NVIDIA RTX 3060 (12 GB)?
Yes, with CPU offloading (rating C, usable).
Which GPUs can run Seedream 3.0?
76 of the 78 GPUs we track can run it, and 15 run it well (rating B or better). The cheapest one that runs it well is the NVIDIA RTX 4070 (12 GB, about $499).
Does Seedream 3.0 work on AMD or Apple Silicon?
19 AMD Radeon and 20 Apple Silicon chips in our list can run it. ROCm support is limited, so AMD may need workarounds. There is no MPS backend, so it may not work on Apple Silicon.