NVIDIA GeForce RTX 3090 Ti × DeepSeek R1 Distill Llama 70B
Can the RTX 3090 Ti run DeepSeek R1 Distill Llama 70B?
No
No. DeepSeek R1 Distill Llama 70B needs at least ~38.9 GB, but the RTX 3090 Ti has ~22.8 GB usable. You would need CPU offloading (much slower) or more memory.
By quantization (8K context)
| Quant | Memory needed | Fits? | Est. tok/s |
|---|---|---|---|
| Q3_K_M | 38.9 GB | ❌ | — |
| Q4_K_M | 47.6 GB | ❌ | — |
| Q5_K_M | 55.3 GB | ❌ | — |
| Q6_K | 63.5 GB | ❌ | — |
| Q8_0 | 80.7 GB | ❌ | — |
| F16 | 149 GB | ❌ | — |
Fastest hardware for DeepSeek R1 Distill Llama 70B
- Apple M5 Max (40-core GPU, 128GB) — ~10.7 tok/s at Q4_K_M
- Apple M2 Ultra (76-core GPU, 192GB) — ~8.4 tok/s at Q4_K_M
- Apple M3 Ultra (80-core GPU, 512GB) — ~8.2 tok/s at Q4_K_M
- Apple M4 Max (40-core GPU, 128GB) — ~7.4 tok/s at Q4_K_M
- Apple M5 Pro (20-core GPU, 64GB) — ~5.9 tok/s at Q4_K_M
Rent a GPU instead
DeepSeek R1 Distill Llama 70B doesn't fit here? Rent a cloud GPU with enough VRAM by the hour and run it there.
Vast.ai — cheap marketplace GPUs · RunPod — on-demand GPU pods
Affiliate links: we may earn a commission at no extra cost to you.
Get new models & benchmarks in your inbox
No spam. Unsubscribe anytime by replying to any email.