DeepSeek-R1-Distill-Llama 70B VRAM: what it needs at Q4_K_M, Q8_0 and FP16
Runs in your browser — nothing you paste leaves this page. How we prove that
LLM VRAM calculator playground
KV cache is counted at FP16. Overhead is a fixed 1.5 GiB estimate for the runtime and compute buffers. A custom size borrows the architecture of the nearest preset.
Results update as you type — press Enter to run now.
GPU tiers
- 8 GiBtoo smallRTX 4060, RTX 3070, RX 7600
- 12 GiBtoo smallRTX 4070, RTX 3060 12GB, RTX 5070
- 16 GiBtoo smallRTX 4060 Ti 16GB, RTX 4080, RTX 5080
- 24 GiBtoo smallRTX 4090, RTX 3090, RX 7900 XTX
- 32 GiBtoo smallRTX 5090, V100 32GB
- 48 GiBfitsRTX A6000, RTX 6000 Ada, L40S
- 80 GiBfitsA100 80GB, H100 80GB
At Q4_K_M with an 8,192-token context, DeepSeek-R1-Distill-Llama 70B needs about 43.8 GiB of VRAM: 39.8 GiB of weights, 2.50 GiB of KV cache and 1.50 GiB of runtime overhead. The smallest common GPU tier that holds it is 48 GiB (RTX A6000, RTX 6000 Ada, L40S).
How much VRAM DeepSeek-R1-Distill-Llama 70B needs
DeepSeek-R1-Distill-Llama 70B, from January 2025, is the largest distill, fine-tuned from Llama-3.3-70B-Instruct, which shares Llama 3.1 70B's architecture. The R1 distills are not DeepSeek-R1 itself, which is a 671B mixture-of-experts model; each one is an existing Qwen2.5 or Llama base fine-tuned on reasoning traces generated by R1, so its memory footprint is exactly that of its base architecture.
With 8,192 tokens of context the estimate is 134 GiB at FP16, 73.3 GiB at Q8_0 and 43.8 GiB at Q4_K_M. FP16 needs more than a single 80 GiB card, Q8_0 needs an 80 GiB card, and Q4_K_M needs a 48 GiB card. The weights are 70 billion parameters times the bits per weight of each format (16, 8.5 and 4.89), divided by eight; everything is in GiB, the unit GPU memory is sold in.
Context length and the KV cache
The KV cache stores a key and a value vector for every token in every layer. DeepSeek-R1-Distill-Llama 70B has 80 layers, 8 KV heads and a head dimension of 128, read from its Hugging Face config.json, so at FP16 each token costs 320 KiB. That is 2.50 GiB at 8,192 tokens, and 40.0 GiB at the maximum of 131,072, where the Q4_K_M total becomes 81.3 GiB and needs more than a single 80 GiB card. The cache does not shrink with weight quantization; only a runtime option that quantizes the cache itself does.
Which quantization to pick
Q4_K_M at 43.8 GiB fits a 48 GiB card or two 24 GiB cards; Q8_0 at 73.3 GiB needs an 80 GiB card.
Pitfalls
Reasoning models write a long chain of thought before the answer, often thousands of tokens, so budget context generously: a reply that runs out of KV cache stops mid-thought. At Q4_K_M, 16,384 tokens need 46.3 GiB and 32,768 need 51.3 GiB, past 48 GiB. On a Mac, the GPU's default share of unified memory is about three quarters, so 64 GB is tight and 96 GB comfortable.
Llama 3.1 70B is the same shape; DeepSeek-R1-Distill-Qwen 32B is the single-24 GiB alternative.
FAQ
Questions, answered.
Tap a question to expand the answer.
Can DeepSeek-R1 70B run on two RTX 3090s?
At Q4_K_M with up to about 16,384 tokens: 43.8 GiB at 8,192 and 46.3 GiB at 16,384, under 48 GiB. At 32,768 tokens it needs 51.3 GiB.
How much VRAM does DeepSeek-R1 70B need at FP16?
134 GiB at 8,192 tokens, more than a single 80 GiB GPU.
More free, private DevOps tools.
The LLM VRAM Calculator is one tool in OpsCanopy — a growing canopy of browser-based validators, converters and testers that never touch a server.
Related tools
New to AI & local LLMs? Read the AI & local LLMs guide →
42 free tools, every one offline-capable — opscanopy.com works with no signup and nothing uploaded.
Related: the full LLM VRAM Calculator, the LLM Token Counter, or every AI tool.
Estimates only; the KV cache is sized at FP16 and the overhead is a fixed allowance, so always confirm against your own runtime’s report before buying hardware. OpsCanopy is free and open.