Run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF locally
unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF is a very large code-focused language model with 30.53 billion parameters, built on the qwen3moe architecture. It is released under the apache-2.0 license and has been downloaded 243,253 times.
To run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF locally at a 4,096-token context, its quantized versions need between 8.44 GB (Q1_0, lowest quality) and 57.89 GB (BF16, highest quality) of memory, weights plus KV cache and a system margin included.
For most users the best balance is Q5_K_XL, needing about 21.23 GB. That means unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF fits entirely in the VRAM of a 10 GB GPU or larger, running fully on the GPU.
All quantizations
| Quant. | Bits | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| Q1_0 | 2.1 | Very low | 7.46 GB | 0.19 GB | 8.44 GB | 53.7 t/s | Fits in VRAM |
| IQ1_S | 2.34 | Very low | 8.3 GB | 0.19 GB | 9.29 GB | 48.2 t/s | Fits in VRAM |
| IQ1_M | 2.52 | Very low | 8.97 GB | 0.19 GB | 9.95 GB | 44.6 t/s | Fits in VRAM |
| IQ2_XXS | 2.71 | Low | 9.62 GB | 0.19 GB | 10.61 GB | 41.6 t/s | Fits in VRAM |
| IQ2_M | 2.84 | Low | 10.09 GB | 0.19 GB | 11.08 GB | 39.6 t/s | Fits in VRAM |
| Q2_K | 2.95 | Low | 10.49 GB | 0.19 GB | 11.47 GB | 38.1 t/s | Fits in VRAM |
| Q2_K_L | 2.97 | Low | 10.55 GB | 0.19 GB | 11.54 GB | 37.9 t/s | Fits in VRAM |
| Q2_K_XL | 3.09 | Low | 10.98 GB | 0.19 GB | 11.97 GB | 36.4 t/s | Fits in VRAM |
| IQ3_XXS | 3.37 | Fair | 11.97 GB | 0.19 GB | 12.95 GB | 33.4 t/s | Fits in VRAM |
| Q3_K_S | 3.48 | Fair | 12.38 GB | 0.19 GB | 13.37 GB | 32.3 t/s | Fits in VRAM |
| Q3_K_XL | 3.62 | Fair | 12.86 GB | 0.19 GB | 13.85 GB | 31.1 t/s | Fits in VRAM |
| Q3_K_M | 3.85 | Fair | 13.7 GB | 0.19 GB | 14.69 GB | 29.2 t/s | Fits in VRAM |
| IQ4_XS | 4.29 | Good | 15.25 GB | 0.19 GB | 16.24 GB | 26.2 t/s | Fits in VRAM |
| IQ4_NL | 4.54 | Good | 16.12 GB | 0.19 GB | 17.11 GB | 24.8 t/s | Fits in VRAM |
| Q4_0 | 4.55 | Good | 16.19 GB | 0.19 GB | 17.17 GB | 24.7 t/s | Fits in VRAM |
| Q4_K_S | 4.57 | Good | 16.26 GB | 0.19 GB | 17.24 GB | 24.6 t/s | Fits in VRAM |
| Q4_K_XL | 4.63 | Good | 16.45 GB | 0.19 GB | 17.44 GB | 24.3 t/s | Fits in VRAM |
| Q4_K_M | 4.86 | Good | 17.28 GB | 0.19 GB | 18.27 GB | 23.1 t/s | Fits in VRAM |
| Q4_1 | 5.03 | Very good | 17.87 GB | 0.19 GB | 18.86 GB | 22.4 t/s | Fits in VRAM |
| Q5_K_S | 5.52 | Very good | 19.63 GB | 0.19 GB | 20.62 GB | 20.4 t/s | Fits in VRAM |
| Q5_K_M | 5.69 | Very good | 20.23 GB | 0.19 GB | 21.22 GB | 19.8 t/s | Fits in VRAM |
| Q5_K_XL | 5.7 | Very good | 20.25 GB | 0.19 GB | 21.23 GB | 19.8 t/s | Fits in VRAM |
| Q6_K | 6.57 | Excellent | 23.37 GB | 0.19 GB | 24.36 GB | 2.1 t/s | Offload |
| Q6_K_XL | 6.9 | Excellent | 24.53 GB | 0.19 GB | 25.52 GB | 2.0 t/s | Offload |
| Q8_0 | 8.51 | Excellent | 30.25 GB | 0.19 GB | 31.24 GB | 1.7 t/s | Offload |
| Q8_K_XL | 9.43 | Excellent | 33.52 GB | 0.19 GB | 34.51 GB | 1.5 t/s | Offload |
| BF16 | 16.01 | Excellent | 56.9 GB | 0.19 GB | 57.89 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-Coder-30B-A3B-Instruct-GGUF?
You need about 9.95 GB of VRAM to run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF entirely on the GPU using the IQ1_M quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.
Can I run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF on an 8 GB GPU?
Partially. unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF only fits on an 8 GB GPU by offloading part of it to system RAM (with Q5_K_XL), which runs but is slower.
Can I run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF on a 16 GB GPU?
Yes. With 16 GB of VRAM you can run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF fully on the GPU using Q3_K_M (about 14.69 GB).
Can I run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF on a 24 GB GPU?
Yes. With 24 GB of VRAM you can run unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF fully on the GPU using Q5_K_XL (about 21.23 GB).
What is the best quantization for unsloth/Qwen3-Coder-30B-A3B-Instruct-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.