// Qwen · Text LLM
Qwen2-72B: VRAM & local deployment
72.7B parameters. About 51.7 GB at 4-bit and 163 GB at BF16 with one 8K-token request.
released 2024-05-22licence: Commercial with conditionsmemory figures are estimates
01Specs
| Publisher | Alibaba Qwen · Qwen |
|---|---|
| Family / series | Qwen · Qwen2 |
| Type | Text LLM (text-generation) |
| Parameters | 72.7B Source: counted from official safetensors weights |
| Context length | 128K tokens (131,072)max_position_embeddings in config.json |
| Architecture | 80 layers · GQA |
| Released weights | BF16 · 145 GBdownload |
| Licence | tongyi-qianwen · Commercial with conditionsCommercial use allowed under the publisher’s own licence terms (user caps, acceptable-use policy, attribution or region limits). Read it before shipping. |
| Released | 2024-05-22 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Total = weights + 10% + 0.5GB overhead + KV cache for one request (FP16 KV). More concurrent users multiply the KV column.
| Precision | Weights | Total @ 8K | Total @ 32K | Total @ 128K | Smallest GPU | Mac |
|---|---|---|---|---|---|---|
| BF16 / FP16 (original) | 145 GB | 163 GB | 171 GB | 203 GB | 4 × RTX PRO 6000 (384GB) | Mac 256GB |
| FP8 / INT8 / Q8_0 | 77.3 GB | 88.2 GB | 96.2 GB | 128 GB | H200 141GB | Mac 256GB |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 44.1 GB | 51.7 GB | 59.7 GB | 91.9 GB | 2 × RTX 5090 (64GB) | Mac 128GB |
KV cache per request: 8K 2.7 GB · 32K 10.7 GB · 128K 43.0 GB
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- Transformers — mentioned in the official model card
- llama.cpp — official GGUF weights
- Ollama — in the official Ollama library · qwen2:72b
ollama run qwen2:72b04Official repositories & quantised builds
6 official repo(s) from Qwen.
- Qwen2-72B main
- Qwen2-72B-Instruct Instruct
- Qwen2-72B-Instruct-GGUF Instruct/GGUF
- Qwen2-72B-Instruct-AWQ Instruct/AWQ
- Qwen2-72B-Instruct-GPTQ-Int8 Instruct/GPTQ/INT8
- Qwen2-72B-Instruct-GPTQ-Int4 Instruct/GPTQ/INT4
05Other Qwen models
All 112 →| Model | Params | 4-bit total | Released |
|---|---|---|---|
| Qwen2-0.5B-Instruct · same series | 494M | 0.9 GB | 2024-06-03 |
| Qwen2-1.5B-Instruct · same series | 1.54B | 1.8 GB | 2024-06-03 |
| Qwen2-7B · same series | 7.62B | 6.0 GB | 2024-06-04 |
| Qwen2-57B-A14B-Instruct · same series | 57.4B / 14B | 39.3 GB | 2024-06-04 |
| Qwen-Image-2.1-Turbo | 7.12B | 5.2 GB | 2026-10-09 |
| Qwen-Image-2.1 | 7.12B | 5.2 GB | 2026-09-14 |
| Qwen3.8-Flash-Next | 180B / 6B | 121 GB | 2026-08-24 |
| Qwen3.8-2.4T-A95B | 2.4T / 95B | 1.60 TB | 2026-08-08 |
| Qwen3.8-27B | 27.8B | 19.6 GB | 2026-08-05 |
| Qwen3.6-27B | 27.8B | 19.6 GB | 2026-04-21 |
| Qwen3.6-35B-A3B | 36B / 3B | 24.6 GB | 2026-04-15 |
| Qwen3.5-0.8B | 873M | 1.2 GB | 2026-02-28 |
06Sources
Data checked 2024-10-09. Parameters, context and licence are copied from the official repository; memory figures are our estimates.