Qwen2.5 32B 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 GiBfitsRTX 4090, RTX 3090, RX 7900 XTX
- 32 GiBfitsRTX 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 32B needs about 21.7 GiB of VRAM: 18.2 GiB of weights, 2.00 GiB of KV cache and 1.50 GiB of runtime overhead. The smallest common GPU tier that holds it is 24 GiB (RTX 4090, RTX 3090, RX 7900 XTX).
How much VRAM Qwen2.5 32B needs
Qwen2.5 32B is the largest Qwen2.5 model that fits a single 24 GiB consumer card when quantized, which makes it a favourite for one-GPU workstations.
With 8,192 tokens of context the estimate is 63.1 GiB at FP16, 35.2 GiB at Q8_0 and 21.7 GiB at Q4_K_M. FP16 needs an 80 GiB card, Q8_0 needs a 48 GiB card, and Q4_K_M needs a 24 GiB card. The weights are 32 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 32B has 64 layers, 8 KV heads and a head dimension of 128, read from its Hugging Face config.json, so at FP16 each token costs 256 KiB. That is 2.00 GiB at 8,192 tokens, and 32.0 GiB at the maximum of 131,072, where the Q4_K_M total becomes 51.7 GiB and needs an 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 21.7 GiB is the 24 GiB answer. Q5_K_M at 24.7 GiB just misses 24 GiB and needs a 32 GiB card; Q8_0 at 35.2 GiB needs a 48 GiB card.
Pitfalls
On a 24 GiB card the margin is thin: 16,384 tokens at Q4_K_M come to 23.7 GiB, and the desktop compositor or a browser using the same GPU can tip it over. Contexts past 32,768 tokens need YaRN rope scaling.
DeepSeek-R1-Distill-Qwen 32B shares the shape; Gemma 2 27B is the other 24 GiB-class choice.
FAQ
Questions, answered.
Tap a question to expand the answer.
Can Qwen2.5 32B run on an RTX 4090?
Yes at Q4_K_M: 21.7 GiB at 8,192 tokens and 23.7 GiB at 16,384, both under 24 GiB. At 32,768 tokens it needs 27.7 GiB.
How much VRAM does Qwen2.5 32B need at Q8_0?
35.2 GiB at 8,192 tokens, so a 48 GiB card or two 24 GiB cards.
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.