Qwen2.5 72B 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, Qwen2.5 72B needs about 45.0 GiB of VRAM: 41.0 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 Qwen2.5 72B needs
Qwen2.5 72B is the top of Alibaba's open Qwen2.5 line and competes with Llama 3.1 70B, with particular strength in maths, code and non-English text.
With 8,192 tokens of context the estimate is 138 GiB at FP16, 75.2 GiB at Q8_0 and 45.0 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 72 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. Qwen2.5 72B 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 82.5 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 45.0 GiB fits a 48 GiB card or a pair of 24 GiB cards. Q5_K_M at 51.8 GiB overflows 48 GiB, and Q8_0 at 75.2 GiB needs an 80 GiB card.
Pitfalls
Its weights are about 1.14 GiB larger than Llama 3.1 70B's at Q4_K_M, which matters when a 48 GiB setup is already full. Native context is 32,768 tokens; longer inputs need YaRN scaling.
Compare Llama 3.1 70B and DeepSeek-R1-Distill-Llama 70B at the same tier, or step down to Qwen2.5 32B.
FAQ
Questions, answered.
Tap a question to expand the answer.
Can Qwen2.5 72B run on 48 GB of VRAM?
At Q4_K_M with 8,192 tokens, yes: 45.0 GiB. At 16,384 tokens the estimate is 47.5 GiB, very close to the limit, and 32,768 tokens need 52.5 GiB.
How much VRAM does Qwen2.5 72B need at FP16?
138 GiB at 8,192 tokens, more than any single 80 GiB card, so FP16 means two or more datacentre GPUs.
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.