RunsOnMy ModelsGPUs & MacsBy VRAMBest GPUCan I run…?API pricesVRAM calculatorNew

Can DeepSeek R1 Distill Llama 70B run on the RTX 4090?

Not fully in VRAM. Even at 2-bit with 8K context, DeepSeek R1 Distill Llama 70B needs about 27.7 GB, more than the 22.8 GB usable on the RTX 4090. It can still run with offload (slower): Q4_K_M with part of the model in system RAM, about 1.1–2.2 tokens/sec (estimate). Details and command.

Runs with offload (slower)

Runs with offload 4-bit (Q4_K_M) at 8K context: 41 of 80 layers on the RTX 4090 (about 22.2 GB) and the other 39 in system RAM (about 20.5 GB). Needs a PC with at least 32 GB of system RAM. Estimated speed: 1.1–2.2 tokens/sec (very slow: fine for batch jobs, painful for chat), versus a bandwidth ceiling of ~24 tok/s if it fit fully in VRAM.

llama-server -hf unsloth/DeepSeek-R1-Distill-Llama-70B-GGUF:Q4_K_M -c 8192 -ngl 41

Estimate, not a benchmark. The command downloads the GGUF from Hugging Face. The speed range spans dual-channel DDR4-3200 to DDR5-6000 system RAM; your CPU, RAM speed and llama.cpp version move it. If llama.cpp runs out of GPU memory, lower -ngl by 2-3; if VRAM is left over, raise it for more speed. Recent llama.cpp builds can also pick the split automatically. How we estimate offload.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •26.5 GB No27.7 GB No35.2 GB No65.2 GB No—
3-bit (Q3_K_M) •33.8 GB No35.1 GB No42.6 GB No72.6 GB No—
4-bit (Q4_K_M) •41.5 GB No42.7 GB No50.2 GB No80.2 GB No—
5-bit (Q5_K_M) •48.4 GB No49.7 GB No57.2 GB No87.2 GB No—
6-bit (Q6_K) •110 GB No111 GB No118 GB No148 GB No—
8-bit (Q8_0) •71.7 GB No73.0 GB No80.5 GB No110 GB No—
16-bit (F16/BF16) •265 GB No266 GB No273 GB No303 GB No—

• real GGUF file size from unsloth/DeepSeek-R1-Distill-Llama-70B-GGUF. Speed is a bandwidth ceiling (1008 GB/s), not a benchmark. Methodology.

Hardware that does fit it

RTX 5090A100 40GBRTX A6000RTX 6000 AdaL40SMac M4 Pro 48GB

What fits instead

"Runs" and "fits" mean the whole model in GPU memory at 8K context (4-bit or better for the model pick). Prices are US launch MSRPs, not today's street prices. "Popular" means the 60 most-downloaded models we track.

More

DeepSeek R1 Distill Llama 70B on every GPU and Mac · Every model that runs on the RTX 4090 · Other models that fit 24 GB VRAM · Try other settings in the calculator

Guides: Best GPU for local LLMs by budget · Can I run gpt-oss, Llama 70B or Qwen3? · Mac vs PC for local LLMs

Data: Hugging Face Hub (deepseek-ai/DeepSeek-R1-Distill-Llama-70B), NVIDIA specs. Model data last checked 2026-11-17; refreshed hourly.