Run unsloth/Qwen3.6-27B-MTP-GGUF locally

License: apache-2.0 ⬇ 868,483 ❤ 874
Parameters27.32B
Context262,144

unsloth/Qwen3.6-27B-MTP-GGUF is a large language model with 27.32 billion parameters, built on the qwen35 architecture. It is released under the apache-2.0 license and has been downloaded 868,483 times.

To run unsloth/Qwen3.6-27B-MTP-GGUF locally at a 4,096-token context, its quantized versions need between 2.14 GB (F16, lowest quality) and 53.05 GB (BF16, highest quality) of memory, weights plus KV cache and a system margin included.

For most users the best balance is F32, needing about 2.99 GB. That means unsloth/Qwen3.6-27B-MTP-GGUF fits entirely in the VRAM of a 6 GB GPU or larger, running fully on the GPU.

→ Guide: How much VRAM do you need?

All quantizations

Quant.Bits QualityWeights KVTotal Speed~Verdict
F16 0.27 Very low 0.86 GB 0.47 GB 2.14 GB 463.0 t/s Fits in VRAM
F32 0.54 Very low 1.72 GB 0.47 GB 2.99 GB 233.0 t/s Fits in VRAM
IQ2_XXS 2.8 Low 8.91 GB 0.47 GB 10.18 GB 5.6 t/s Offload
IQ2_M 3.23 Low 10.27 GB 0.47 GB 11.55 GB 4.9 t/s Offload
Q2_K_XL 3.53 Fair 11.21 GB 0.47 GB 12.49 GB 4.5 t/s Offload
IQ3_XXS 3.57 Fair 11.37 GB 0.47 GB 12.64 GB 4.4 t/s Offload
Q3_K_S 3.68 Fair 11.71 GB 0.47 GB 12.99 GB 4.3 t/s Offload
Q3_K_M 4.05 Fair 12.87 GB 0.47 GB 14.14 GB 3.9 t/s Offload
Q3_K_XL 4.33 Good 13.77 GB 0.47 GB 15.05 GB 3.6 t/s Offload
IQ4_XS 4.6 Good 14.63 GB 0.47 GB 15.9 GB 3.4 t/s Offload
Q4_0 4.7 Good 14.95 GB 0.47 GB 16.23 GB 3.3 t/s Offload
Q4_K_S 4.72 Good 15.01 GB 0.47 GB 16.29 GB 3.3 t/s Offload
IQ4_NL 4.78 Good 15.22 GB 0.47 GB 16.49 GB 3.3 t/s Offload
Q4_K_M 5.01 Very good 15.93 GB 0.47 GB 17.21 GB 3.1 t/s Offload
Q4_1 5.14 Very good 16.34 GB 0.47 GB 17.61 GB 3.1 t/s Offload
Q4_K_XL 5.24 Very good 16.68 GB 0.47 GB 17.95 GB 3.0 t/s Offload
Q5_K_S 5.64 Very good 17.95 GB 0.47 GB 19.22 GB 2.8 t/s Offload
Q5_K_M 5.81 Very good 18.47 GB 0.47 GB 19.75 GB 2.7 t/s Offload
Q5_K_XL 5.96 Very good 18.95 GB 0.47 GB 20.23 GB 2.6 t/s Offload
Q6_K 6.7 Excellent 21.31 GB 0.47 GB 22.59 GB 2.3 t/s Offload
Q6_K_XL 7.62 Excellent 24.23 GB 0.47 GB 25.5 GB Insufficient
Q8_0 8.51 Excellent 27.05 GB 0.47 GB 28.33 GB Insufficient
Q8_K_XL 10.48 Excellent 33.32 GB 0.47 GB 34.59 GB Insufficient
BF16 16.28 Excellent 51.77 GB 0.47 GB 53.05 GB Insufficient

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

Frequently asked questions

How much VRAM do you need to run unsloth/Qwen3.6-27B-MTP-GGUF?

You need about 2.99 GB of VRAM to run unsloth/Qwen3.6-27B-MTP-GGUF entirely on the GPU using the F32 quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.

Can I run unsloth/Qwen3.6-27B-MTP-GGUF on an 8 GB GPU?

Yes. With 8 GB of VRAM you can run unsloth/Qwen3.6-27B-MTP-GGUF fully on the GPU using F32 (about 2.99 GB).

Can I run unsloth/Qwen3.6-27B-MTP-GGUF on a 16 GB GPU?

Yes. With 16 GB of VRAM you can run unsloth/Qwen3.6-27B-MTP-GGUF fully on the GPU using IQ4_XS (about 15.9 GB).

Can I run unsloth/Qwen3.6-27B-MTP-GGUF on a 24 GB GPU?

Yes. With 24 GB of VRAM you can run unsloth/Qwen3.6-27B-MTP-GGUF fully on the GPU using Q6_K (about 22.59 GB).

What is the best quantization for unsloth/Qwen3.6-27B-MTP-GGUF?

If memory allows, higher bits-per-weight means better quality. A common sweet spot is a Q4_K_M or Q5_K_M quantization, which keeps most of the quality while roughly halving the memory versus 8-bit. Pick the highest quantization that still fits in your VRAM.