Qwen3-235B-A22B GGUF size and VRAM requirements
unsloth/Qwen3-235B-A22B-GGUF is a very large language model with 235.09 billion parameters, built on the qwen3moe architecture. It is released under the apache-2.0 license and has been downloaded 47,897 times.
To run unsloth/Qwen3-235B-A22B-GGUF locally at a 4,096-token context, its quantized versions need between 81.34 GB (Q2_K, lowest quality) and 439.53 GB (BF16, highest quality) of memory, weights plus KV cache and a system margin included.
Available GGUF quantizations for unsloth/Qwen3-235B-A22B-GGUF include Q2_K, Q2_K_L, Q2_K_XL, Q3_K_S, Q3_K_XL, Q3_K_M, IQ4_XS, Q4_K_S, Q4_K_XL, Q4_K_M, Q4_1, Q5_K_S, Q5_K_M, Q5_K_XL, Q6_K, Q6_K_XL, Q8_0, Q8_K_XL, BF16. The model supports a native context length of up to 40,960 tokens; a longer context grows the KV cache and the memory needed.
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 | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| Q2_K | 2.92 | Low | 79.81 GB | 0.73 GB | 81.34 GB | — | Insufficient |
| Q2_K_L | 2.92 | Low | 79.94 GB | 0.73 GB | 81.48 GB | — | Insufficient |
| Q2_K_XL | 3.0 | Low | 81.97 GB | 0.73 GB | 83.5 GB | — | Insufficient |
| Q3_K_S | 3.45 | Fair | 94.48 GB | 0.73 GB | 96.01 GB | — | Insufficient |
| Q3_K_XL | 3.53 | Fair | 96.6 GB | 0.73 GB | 98.13 GB | — | Insufficient |
| Q3_K_M | 3.83 | Fair | 104.72 GB | 0.73 GB | 106.26 GB | — | Insufficient |
| IQ4_XS | 4.27 | Good | 116.89 GB | 0.73 GB | 118.42 GB | — | Insufficient |
| Q4_K_S | 4.55 | Good | 124.51 GB | 0.73 GB | 126.04 GB | — | Insufficient |
| Q4_K_XL | 4.56 | Good | 124.91 GB | 0.73 GB | 126.45 GB | — | Insufficient |
| Q4_K_M | 4.84 | Good | 132.39 GB | 0.73 GB | 133.93 GB | — | Insufficient |
| Q4_1 | 5.01 | Very good | 137.12 GB | 0.73 GB | 138.65 GB | — | Insufficient |
| Q5_K_S | 5.51 | Very good | 150.76 GB | 0.73 GB | 152.3 GB | — | Insufficient |
| Q5_K_M | 5.68 | Very good | 155.36 GB | 0.73 GB | 156.89 GB | — | Insufficient |
| Q5_K_XL | 5.68 | Very good | 155.42 GB | 0.73 GB | 156.96 GB | — | Insufficient |
| Q6_K | 6.57 | Excellent | 179.76 GB | 0.73 GB | 181.29 GB | — | Insufficient |
| Q6_K_XL | 6.77 | Excellent | 185.2 GB | 0.73 GB | 186.73 GB | — | Insufficient |
| Q8_0 | 8.51 | Excellent | 232.77 GB | 0.73 GB | 234.31 GB | — | Insufficient |
| Q8_K_XL | 9.02 | Excellent | 246.88 GB | 0.73 GB | 248.41 GB | — | Insufficient |
| BF16 | 16.0 | Excellent | 437.99 GB | 0.73 GB | 439.53 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 unsloth/Qwen3-235B-A22B-GGUF?
unsloth/Qwen3-235B-A22B-GGUF is a language model with 235.09 billion parameters, based on the qwen3moe architecture. It is released under the apache-2.0 license and distributed as GGUF files for local inference.
Can I run unsloth/Qwen3-235B-A22B-GGUF on an 8 GB GPU?
No. unsloth/Qwen3-235B-A22B-GGUF does not fit on an 8 GB GPU, even with the smallest quantization and system RAM offloading.
Can I run unsloth/Qwen3-235B-A22B-GGUF on a 16 GB GPU?
No. unsloth/Qwen3-235B-A22B-GGUF does not fit on a 16 GB GPU, even with the smallest quantization and system RAM offloading.
Can I run unsloth/Qwen3-235B-A22B-GGUF on a 24 GB GPU?
No. unsloth/Qwen3-235B-A22B-GGUF does not fit on a 24 GB GPU, even with the smallest quantization and system RAM offloading.
What context length does unsloth/Qwen3-235B-A22B-GGUF support?
unsloth/Qwen3-235B-A22B-GGUF supports a native context length of up to 40,960 tokens. A longer context grows the KV cache, so it increases the memory needed to run the model.
What is the best quantization for unsloth/Qwen3-235B-A22B-GGUF?
For unsloth/Qwen3-235B-A22B-GGUF, a strong default is Q4_K_M, which needs about 133.93 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.