Qwen2.5-Coder-32B-Instruct GGUF size and VRAM requirements

License: apache-2.0 ⬇ 104,427 ❤ 120
Parameters32.76B
Context32,768

Qwen2.5-Coder-32B-Instruct is a 32.76 billion-parameter AI model licensed under Apache 2.0. It is part of the Qwen family, designed for coding-related tasks such as code generation and understanding. The model is optimized for instruction-following and is available for use under the specified license.

Under the hood it uses 64 transformer layers, a hidden size of 5,120, 40 attention heads. It uses grouped-query attention (40 query heads sharing 8 key/value heads), which already trims KV-cache memory compared with full multi-head attention.

To run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF locally at a 4,096-token context, its quantized versions need between 10.21 GB (IQ2_XXS, lowest quality) and 34.23 GB (Q8_0, highest quality) of memory, weights plus KV cache and a system margin included. Context length drives that memory directly: at 4,096 tokens the KV cache for Qwen2.5-Coder-32B-Instruct-GGUF is about 1.0 GB, rising to roughly 8.0 GB at its full 32,768-token context. Shorter prompts free up memory for a higher-quality quantization.

For most users the best balance is Q5_K_L, needing about 23.91 GB. That means bartowski/Qwen2.5-Coder-32B-Instruct-GGUF fits entirely in the VRAM of a 12 GB GPU or larger, running fully on the GPU.

Available GGUF quantizations for bartowski/Qwen2.5-Coder-32B-Instruct-GGUF include IQ2_XXS, IQ2_XS, IQ2_S, IQ2_M, Q2_K, IQ3_XXS, Q2_K_L, IQ3_XS, Q3_K_S, IQ3_M, Q3_K_M, Q3_K_L, IQ4_XS, Q3_K_XL, Q4_0_4_4, Q4_0_4_8, Q4_0_8_8, IQ4_NL, Q4_0, Q4_K_S, Q4_K_M, Q4_K_L, Q5_K_S, Q5_K_M, Q5_K_L, Q6_K, Q6_K_L, Q8_0. The model supports a native context length of up to 32,768 tokens; a longer context grows the KV cache and the memory needed.

→ Guide: How much VRAM do you need?

GGUF file size and memory by quantization

Compare real GGUF weight sizes, estimated KV cache and total memory for Q4, Q5, Q8 and every quantization published in this repository.

Quant.Bits QualityWeights KVTotal Speed~Verdict
IQ2_XXS 2.2 Very low 8.41 GB 1.0 GB 10.21 GB 5.9 t/s Offload
IQ2_XS 2.43 Very low 9.27 GB 1.0 GB 11.07 GB 5.4 t/s Offload
IQ2_S 2.54 Very low 9.67 GB 1.0 GB 11.47 GB 5.2 t/s Offload
IQ2_M 2.75 Low 10.49 GB 1.0 GB 12.29 GB 4.8 t/s Offload
Q2_K 3.01 Low 11.47 GB 1.0 GB 13.27 GB 4.4 t/s Offload
IQ3_XXS 3.13 Low 11.96 GB 1.0 GB 13.76 GB 4.2 t/s Offload
Q2_K_L 3.19 Low 12.18 GB 1.0 GB 13.98 GB 4.1 t/s Offload
IQ3_XS 3.35 Fair 12.76 GB 1.0 GB 14.56 GB 3.9 t/s Offload
Q3_K_S 3.51 Fair 13.4 GB 1.0 GB 15.2 GB 3.7 t/s Offload
IQ3_M 3.62 Fair 13.79 GB 1.0 GB 15.59 GB 3.6 t/s Offload
Q3_K_M 3.89 Fair 14.84 GB 1.0 GB 16.64 GB 3.4 t/s Offload
Q3_K_L 4.21 Good 16.06 GB 1.0 GB 17.86 GB 3.1 t/s Offload
IQ4_XS 4.32 Good 16.48 GB 1.0 GB 18.28 GB 3.0 t/s Offload
Q3_K_XL 4.38 Good 16.7 GB 1.0 GB 18.5 GB 3.0 t/s Offload
Q4_0_4_4 4.55 Good 17.36 GB 1.0 GB 19.16 GB 2.9 t/s Offload
Q4_0_4_8 4.55 Good 17.36 GB 1.0 GB 19.16 GB 2.9 t/s Offload
Q4_0_8_8 4.55 Good 17.36 GB 1.0 GB 19.16 GB 2.9 t/s Offload
IQ4_NL 4.56 Good 17.4 GB 1.0 GB 19.2 GB 2.9 t/s Offload
Q4_0 4.57 Good 17.43 GB 1.0 GB 19.23 GB 2.9 t/s Offload
Q4_K_S 4.59 Good 17.49 GB 1.0 GB 19.29 GB 2.9 t/s Offload
Q4_K_M 4.85 Good 18.49 GB 1.0 GB 20.29 GB 2.7 t/s Offload
Q4_K_L 4.99 Good 19.03 GB 1.0 GB 20.83 GB 2.6 t/s Offload
Q5_K_S 5.53 Very good 21.08 GB 1.0 GB 22.88 GB 2.4 t/s Offload
Q5_K_M 5.68 Very good 21.66 GB 1.0 GB 23.46 GB 2.3 t/s Offload
Q5_K_L 5.8 Very good 22.11 GB 1.0 GB 23.91 GB 2.3 t/s Offload
Q6_K 6.56 Excellent 25.04 GB 1.0 GB 26.84 GB Insufficient
Q6_K_L 6.66 Excellent 25.39 GB 1.0 GB 27.19 GB Insufficient
Q8_0 8.5 Excellent 32.43 GB 1.0 GB 34.23 GB Insufficient

KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.

Frequently asked questions

What kind of model is bartowski/Qwen2.5-Coder-32B-Instruct-GGUF?

bartowski/Qwen2.5-Coder-32B-Instruct-GGUF is a code-focused language model with 32.76 billion parameters, based on the qwen2 architecture. It is released under the apache-2.0 license and distributed as GGUF files for local inference.

How much VRAM do you need to run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF?

You need about 11.47 GB of VRAM to run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF entirely on the GPU using the IQ2_S quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.

Can I run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF on an 8 GB GPU?

Partially. bartowski/Qwen2.5-Coder-32B-Instruct-GGUF only fits on an 8 GB GPU by offloading part of it to system RAM (with Q5_K_L), which runs but is slower.

Can I run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF on a 16 GB GPU?

Yes. With 16 GB of VRAM you can run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF fully on the GPU using IQ3_M (about 15.59 GB).

Can I run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF on a 24 GB GPU?

Yes. With 24 GB of VRAM you can run bartowski/Qwen2.5-Coder-32B-Instruct-GGUF fully on the GPU using Q5_K_L (about 23.91 GB).

What context length does bartowski/Qwen2.5-Coder-32B-Instruct-GGUF support?

bartowski/Qwen2.5-Coder-32B-Instruct-GGUF supports a native context length of up to 32,768 tokens. A longer context grows the KV cache, so it increases the memory needed to run the model.

What is the best quantization for bartowski/Qwen2.5-Coder-32B-Instruct-GGUF?

For bartowski/Qwen2.5-Coder-32B-Instruct-GGUF, a strong default is Q4_K_M, which needs about 20.29 GB and keeps most of the quality while roughly halving the memory versus 8-bit. With VRAM to spare, Q5_K_M or Q6_K add a little more quality; if you are tight on memory, a smaller quantization still runs. Pick the highest quantization that fits your VRAM.