DeepSeek-R1-Distill-Llama-70B
70BDeepSeek
R1 reasoning in Llama 70B architecture. Top open reasoning model for dual-GPU setups.
131K
Max Context
2
Quant Variants
GGUF Q4_K_M
Best Quality
97.6%
Accuracy Retained
Quantization Variants
Per-quant VRAM, quality loss, and inference speed on RTX 4090
Measured = site benchmarks · Estimated = formula · Community = public reports
Similar models
Compare with DeepSeek-V3DeepSeek-V3
DeepSeek
DeepSeek-V3 frontier MoE (~37B active / 671B total). MLA + FP8; multi-node GPU cluster required at Q4.
DeepSeek-R1
DeepSeek
DeepSeek-R1 reasoning model built on V3 MoE. Chain-of-thought at frontier scale — use distill variants for local GPUs.
DeepSeek-R1-Distill-Qwen-32B
DeepSeek
R1 distilled to 32B. Near-frontier reasoning on a single 24GB card (Q3/Q4).
DeepSeek-R1-Distill-Qwen-14B
DeepSeek
R1 reasoning distilled into 14B. Huge community interest; excellent chain-of-thought.
Running DeepSeek-R1-Distill-Llama-70B locally
At Q4_K_M and 4K of context, DeepSeek-R1-Distill-Llama-70B needs about 46.1 GB — 40.7 GB of weights, 1.25 GB of KV cache and a 4.2 GB activation buffer. The smallest card in this index that clears that comfortably is the 64 GB RAM (CPU) at 64 GB, and 12 of the 63 cards here do. These are calculated figures, not measurements: the estimate stops counting a card as comfortable at 88% of its VRAM, which is roughly the room a desktop session needs.
What longer context costs
Going from 4K to 32K adds about 8.75 GB, taking the total to 55.7 GB. Weights do not move with context — only the KV cache does, and it grows linearly, so this is the number to watch when planning for long documents. The model's native window is 128K; holding all of it at Q4_K_M would need about 89 GB.
Which build to download
This index tracks 2 formats for it — GGUF, AWQ — across 2 levels. Q4_K_M carries the lowest published perplexity loss at 2.4%. The fastest level measured here is AWQ INT4 at 52 tok/s on an RTX 4090, batch 1. GGUF runs on llama.cpp and Ollama across NVIDIA, AMD and Apple silicon; AWQ and GPTQ target vLLM on CUDA and ROCm; EXL2 is ExLlamaV2 and CUDA only.
Common questions
- How much VRAM does DeepSeek-R1-Distill-Llama-70B need?
- About 46.1 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 55.7 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.
- Will DeepSeek-R1-Distill-Llama-70B run on a 64GB GPU?
- Yes — at Q4_K_M and 4K context it needs about 46.1 GB, which leaves 17.9 GB spare on a 64 GB RAM (CPU). That is the smallest card in this index that clears it comfortably; 12 of 63 do.
- Which quantization of DeepSeek-R1-Distill-Llama-70B should I use?
- Q4_K_M has the lowest published quality loss (2.4%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, AWQ INT4.
Where this model fits
Sized at Q4_K_M with a 4K context window, smallest card first. Comfortable means the estimate uses at most 88% of the memory.
- Runs comfortably on
- A100 80G80GB · +33.9H100 80G80GB · +33.9Mac M2 Max 96G96GB · +25.9Mac M5 Max 128G128GB · +49.9Mac M4 Max 128G128GB · +49.9
- Tight but possible
- L40S 48G48GBA40 48G48GBRadeon PRO W7900 48G48GB
- Compared here
- GGUF vs AWQ
- Try it yourself
- Size it in the calculatorGenerate the run command
How to actually run this
Deployment guides for this model and this class of hardware.
This page's figures change when the model or the runtime does.Last updated 2026-09-21 RSS → /feed.xml