Qwen_Qwen3-Coder-Next GGUF size and VRAM requirements

License: apache-2.0 ⬇ 31,139 ❤ 25
Parameters79.67B
Context262,144

bartowski/Qwen_Qwen3-Coder-Next-GGUF is a very large code-focused language model with 79.67 billion parameters, built on the qwen3next architecture. It is released under the apache-2.0 license and has been downloaded 31,139 times.

Qwen_Qwen3-Coder-Next-GGUF is a Mixture-of-Experts model with 512 experts, of which 10 are active on each token. Routing only 10 of 512 experts makes it noticeably faster than a dense model of the same size — but every expert still has to be held in memory, so the numbers below are set by the full parameter count, not the active one. It is built from 48 transformer layers, a hidden size of 2,048, 16 attention heads. It uses grouped-query attention (16 query heads sharing 2 key/value heads), which already trims KV-cache memory compared with full multi-head attention.

To run bartowski/Qwen_Qwen3-Coder-Next-GGUF locally at a 4,096-token context, its quantized versions need between 16.62 GB (IQ1_S, lowest quality) and 80.16 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 Qwen_Qwen3-Coder-Next-GGUF is about 0.38 GB, rising to roughly 24.0 GB at its full 262,144-token context. Shorter prompts free up memory for a higher-quality quantization.

For most users the best balance is IQ2_S, needing about 22.94 GB. That means bartowski/Qwen_Qwen3-Coder-Next-GGUF fits entirely in the VRAM of a 24 GB GPU or larger, running fully on the GPU.

Available GGUF quantizations for bartowski/Qwen_Qwen3-Coder-Next-GGUF include IQ1_S, IQ1_M, IQ2_XXS, IQ2_XS, IQ2_S, IQ2_M, Q2_K, Q2_K_L, IQ3_XXS, IQ3_XS, Q3_K_S, IQ3_M, Q3_K_M, Q3_K_L, Q3_K_XL, IQ4_XS, IQ4_NL, Q4_0, Q4_K_S, Q4_K_M, Q4_K_L, Q4_1, 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 262,144 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
IQ1_S 1.67 Very low 15.44 GB 0.38 GB 16.62 GB 3.2 t/s Offload
IQ1_M 1.74 Very low 16.11 GB 0.38 GB 17.29 GB 3.1 t/s Offload
IQ2_XXS 1.94 Very low 17.97 GB 0.38 GB 19.15 GB 2.8 t/s Offload
IQ2_XS 2.23 Very low 20.69 GB 0.38 GB 21.87 GB 2.4 t/s Offload
IQ2_S 2.35 Very low 21.76 GB 0.38 GB 22.94 GB 2.3 t/s Offload
IQ2_M 2.62 Low 24.31 GB 0.38 GB 25.48 GB Insufficient
Q2_K 2.83 Low 26.23 GB 0.38 GB 27.41 GB Insufficient
Q2_K_L 2.86 Low 26.52 GB 0.38 GB 27.69 GB Insufficient
IQ3_XXS 3.19 Low 29.55 GB 0.38 GB 30.72 GB Insufficient
IQ3_XS 3.32 Fair 30.76 GB 0.38 GB 31.94 GB Insufficient
Q3_K_S 3.5 Fair 32.47 GB 0.38 GB 33.64 GB Insufficient
IQ3_M 3.68 Fair 34.13 GB 0.38 GB 35.3 GB Insufficient
Q3_K_M 3.68 Fair 34.14 GB 0.38 GB 35.31 GB Insufficient
Q3_K_L 3.84 Fair 35.6 GB 0.38 GB 36.78 GB Insufficient
Q3_K_XL 3.87 Fair 35.86 GB 0.38 GB 37.03 GB Insufficient
IQ4_XS 4.3 Good 39.91 GB 0.38 GB 41.08 GB Insufficient
IQ4_NL 4.55 Good 42.2 GB 0.38 GB 43.37 GB Insufficient
Q4_0 4.63 Good 42.93 GB 0.38 GB 44.11 GB Insufficient
Q4_K_S 4.71 Good 43.71 GB 0.38 GB 44.88 GB Insufficient
Q4_K_M 4.89 Good 45.38 GB 0.38 GB 46.56 GB Insufficient
Q4_K_L 4.92 Good 45.6 GB 0.38 GB 46.77 GB Insufficient
Q4_1 5.04 Very good 46.78 GB 0.38 GB 47.95 GB Insufficient
Q5_K_S 5.54 Very good 51.38 GB 0.38 GB 52.56 GB Insufficient
Q5_K_M 5.72 Very good 53.07 GB 0.38 GB 54.24 GB Insufficient
Q5_K_L 5.74 Very good 53.25 GB 0.38 GB 54.42 GB Insufficient
Q6_K 6.61 Excellent 61.27 GB 0.38 GB 62.45 GB Insufficient
Q6_K_L 6.62 Excellent 61.41 GB 0.38 GB 62.59 GB Insufficient
Q8_0 8.52 Excellent 78.99 GB 0.38 GB 80.16 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/Qwen_Qwen3-Coder-Next-GGUF?

bartowski/Qwen_Qwen3-Coder-Next-GGUF is a code-focused language model with 79.67 billion parameters, based on the qwen3next architecture. It is released under the apache-2.0 license and distributed as GGUF files for local inference.

Is bartowski/Qwen_Qwen3-Coder-Next-GGUF a Mixture-of-Experts (MoE) model?

Yes. bartowski/Qwen_Qwen3-Coder-Next-GGUF is a Mixture-of-Experts model with 512 experts, of which 10 are activated per token. That makes it faster than a dense model of the same size, but all 512 experts must be loaded into memory, so the VRAM/RAM it needs is driven by the total parameter count, not the active one.

How much VRAM do you need to run bartowski/Qwen_Qwen3-Coder-Next-GGUF?

You need about 22.94 GB of VRAM to run bartowski/Qwen_Qwen3-Coder-Next-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/Qwen_Qwen3-Coder-Next-GGUF on an 8 GB GPU?

Partially. bartowski/Qwen_Qwen3-Coder-Next-GGUF only fits on an 8 GB GPU by offloading part of it to system RAM (with IQ2_S), which runs but is slower.

Can I run bartowski/Qwen_Qwen3-Coder-Next-GGUF on a 16 GB GPU?

Partially. bartowski/Qwen_Qwen3-Coder-Next-GGUF only fits on a 16 GB GPU by offloading part of it to system RAM (with Q4_1), which runs but is slower.

Can I run bartowski/Qwen_Qwen3-Coder-Next-GGUF on a 24 GB GPU?

Yes. With 24 GB of VRAM you can run bartowski/Qwen_Qwen3-Coder-Next-GGUF fully on the GPU using IQ2_S (about 22.94 GB).

What context length does bartowski/Qwen_Qwen3-Coder-Next-GGUF support?

bartowski/Qwen_Qwen3-Coder-Next-GGUF supports a native context length of up to 262,144 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/Qwen_Qwen3-Coder-Next-GGUF?

For bartowski/Qwen_Qwen3-Coder-Next-GGUF, a strong default is Q4_K_M, which needs about 46.56 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.