Can I Run DeepSeek-R1-Distill-Qwen-32B on 96 GB system RAM?
Written by Jakub Rusinowski · Last updated January 20, 2025
These figures are for DeepSeek-R1-Distill-Qwen-32B, a distill of Qwen2.5-32B — not the full DeepSeek R1. The full DeepSeek R1 (671B) needs about 405 GB of weights at Q4_K_M and is a different model.
Yes — comfortably
Yes, comfortably — DeepSeek R1 Distill Qwen 32B at Q2_K needs about 13.5 GB of the 76.8 GB usable on 96 GB system RAM, leaving ~63.3 GB spare and running at ~5.7 tok/s (estimated), with room for about 131,072 tokens of context.
Confidence: high · Recommended quantization: Q2_K · Estimated speed: ~5.7 tok/s
See what else this hardware can run →
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
96 GB system RAM — what it gives a model
| Usable memory for models | 76.8 GB |
| Memory bandwidth | 90 GB/s |
DeepSeek R1 on 96 GB system RAM: memory by quantization
| Quant | Memory needed | Fits 76.8 GB? | Max context | Est. speed | Download |
|---|---|---|---|---|---|
| F16 | 66.9 GB | ✓ Yes | 32K | ~1 tok/s | 64 GB |
| Q8_0 | 36.9 GB | ✓ Yes | 128K | ~1.9 tok/s | 34 GB |
| Q6_K | 29.2 GB | ✓ Yes | 128K | ~2.4 tok/s | 26.2 GB |
| Q5_K_M | 25.6 GB | ✓ Yes | 128K | ~2.8 tok/s | 22.7 GB |
| Q4_K_M | 22.3 GB | ✓ Yes | 128K | ~3.3 tok/s | 19.3 GB |
| Q3_K_M | 16.6 GB | ✓ Yes | 128K | ~4.5 tok/s | 13.6 GB |
| Q2_K | 13.5 GB | ✓ Yes | 128K | ~5.7 tok/s | 10.5 GB |
Which DeepSeek R1 sizes fit
| Variant | Needs at Q4_K_M | Fits? | Est. speed |
|---|---|---|---|
| DeepSeek R1 (671B) | 406.5 GB | ✗ Too large | — |
| DeepSeek R1 Distill Qwen 32B | 22.3 GB | ✓ Fits | ~3.3 tok/s |
| DeepSeek R1 Distill Qwen 14B | 10.6 GB | ✓ Fits | ~7.2 tok/s |
| DeepSeek R1 Distill Llama 8B | 6.7 GB | ✓ Fits | ~12.2 tok/s |
What to watch out for
- Q2_K is the only quantization that fits, and it is a heavily degraded one — expect noticeably worse output than the same model at Q4_K_M. A smaller model at Q4 is usually the better trade.
- 1 larger variant of DeepSeek R1 does not fit and would need CPU offload or different hardware.
- These figures assume CPU-only inference. Any discrete GPU, even an 8 GB one, will be several times faster for models that fit in its VRAM.
Recommended setup
llama.cpp (CPU build) or Ollama — both run without a GPU
How these numbers are calculated
- Memory need = quantized weights + KV cache at 8,192 tokens (f16) + 0.8 GB runtime overhead.
- 76.8 GB of the 96 GB is treated as usable for model weights (80% — the rest is the OS and running applications).
- DDR5-5600 dual channel at 89.6 GB/s peak. CPU decode is assumed to sustain 35% of that peak, because CPU inference is not purely bandwidth-bound — it also spends real time in compute and thread synchronisation. This figure is an assumption, not a fitted constant: no CPU measurement is in the calibration set.
- CPU-only inference: no GPU is assumed. A GPU of any size will beat these figures substantially.
- Throughput is a memory-bandwidth roofline estimate, not a measurement. It is labelled "estimated" everywhere it appears.
- KV cache is computed from this model's published attention configuration.
FAQ
Can I run DeepSeek R1 on 96 GB system RAM?
Yes, comfortably — DeepSeek R1 Distill Qwen 32B at Q2_K needs about 13.5 GB of the 76.8 GB usable on 96 GB system RAM, leaving ~63.3 GB spare and running at ~5.7 tok/s (estimated), with room for about 131,072 tokens of context.
Which quantization of DeepSeek R1 should I use on 96 GB system RAM?
Q2_K — it needs about 13.5 GB of the 76.8 GB available, downloads as roughly 10.5 GB, and runs at an estimated 5.7 tokens/sec with up to 128K of context.
What limits DeepSeek R1 on 96 GB system RAM?
Nothing binding — the model fits with headroom and generates at a usable speed on this hardware.
Which runtime should I use?
llama.cpp (CPU build) or Ollama — both run without a GPU
Other Models on 96 GB system RAM
- Devstral on 96 GB system RAM
- EXAONE 3.5 on 96 GB system RAM
- Falcon 3 on 96 GB system RAM
- Gemma 2 Family on 96 GB system RAM
- Gemma 3 on 96 GB system RAM
DeepSeek R1 on GPUs
- DeepSeek R1 on NVIDIA GeForce RTX 5090
- DeepSeek R1 on NVIDIA GeForce RTX 5080
- DeepSeek R1 on NVIDIA GeForce RTX 5070 Ti
- DeepSeek R1 on NVIDIA GeForce RTX 5070
What This Model Is Good At
Model & Tools
← Can I Run It? | DeepSeek R1 model page | Check your hardware