Qwen3 Coder 30B A3B, self-quantized to GGUF by Atomic Chat. Built straight from Qwen's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline.
Highlights
- 30.5B parameters: the weights this repo quantizes.
- Context length: 262,144 tokens (256K), as published by Qwen.
- 48 layers: Mixture-of-Experts.
- Full imatrix ladder: every quant is calibrated with an importance matrix.
- Significant Performance: among open models on Agentic Coding, Agentic Browser-Use, and other foundational coding tasks.
- Agentic Coding: supporting for most platform such as Qwen Code, CLINE, featuring a specially designed function call format.
[!NOTE] These GGUFs are self-quantized from the original weights, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model.
[!IMPORTANT] Always pass
--jinjaso the Qwen3 Coder 30B A3B chat template is applied. Without it the model can emit malformed turns.
Model Overview
| Property | Value |
|---|---|
| Base model | Qwen/Qwen3-Coder-30B-A3B-Instruct |
| Parameters | 30.5B |
| Layers | 48 |
| Experts | 128 routed (top-8) |
| Context length | 262,144 tokens (256K) |
| Vocabulary | 151,936 |
| Modalities | Text |
| Architecture | Mixture-of-Experts, 128 experts (top-8), 32 attention heads over 4 KV heads, Qwen3MoeForCausalLM |
| This repo | GGUF quants (imatrix). Quants: Q2_K, Q5_K_M, IQ3_M, Q3_K_M, Q3_K_L, IQ4_XS, Q6_K, Q4_K_S, Q4_K_M, UD-Q4_K_XL, Q5_K_S, Q8_0 |
Choosing a quant
| Quant | Size | Notes |
|---|---|---|
Q2_K | 11.3 GB | Smallest K-quant. Minimal RAM, clear quality drop. |
Q5_K_M | 12.1 GB | Higher quality, low loss. |
IQ3_M | 13.5 GB | Beats Q3 at a similar size thanks to imatrix. Best low-RAM pick. |
Q3_K_M | 14.7 GB | Low quality but usable. |
Q3_K_L | 15.9 GB | A step above Q3_K_M. |
IQ4_XS | 16.4 GB | Excellent quality for size. Recommended low-bit. |
Q6_K | 17.4 GB | Near lossless, noticeably lighter than Q8_0. |
Q4_K_S | 17.5 GB | Compact 4-bit, fast. |
Q4_K_M | 18.6 GB | Recommended default. Best balance of size, speed and quality. |
UD-Q4_K_XL | 18.8 GB | Dynamic. Embeddings and output kept at Q8_0 for higher quality at a Q4 footprint. |
Q5_K_S | 19.7 GB | Higher quality, slightly more compact than Q5_K_M. |
Q8_0 | 20.3 GB | Effectively lossless, reference quality. |
[!TIP] Pick the largest file that fits your (V)RAM with room for context.
Q4_K_MorUD-Q4_K_XLis the sweet spot for most setups;Q6_KorQ8_0for maximum fidelity.
Get started
Run Qwen3 Coder 30B A3B locally with:
- Atomic Chat: the easiest path. Open the app, search
AtomicChat/qwen3-coder-30b-a3b-GGUF, pick a quant, hit Use this model. - llama.cpp:
llama-server -hf AtomicChat/qwen3-coder-30b-a3b-GGUF:Q4_K_M --jinja -c 8192 - Ollama:
ollama run hf.co/AtomicChat/qwen3-coder-30b-a3b-GGUF:Q4_K_M - LM Studio / Jan: search the repo id, download any quant.
Best practices
| Parameter | Value |
|---|---|
| temperature | 0.7 |
| top_p | 0.8 |
| top_k | 20 |
| repetition_penalty | 1.05 |
Qwen's recommended sampling configuration for Qwen/Qwen3-Coder-30B-A3B-Instruct.
Run in llama.cpp
git clone https://github.com/ggml-org/llama.cpp
cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=ON
cmake --build llama.cpp/build --config Release -j --target llama-cli llama-server
./llama.cpp/build/bin/llama-server \
-hf AtomicChat/qwen3-coder-30b-a3b-GGUF:Q4_K_M \
--jinja -ngl 99 -c 8192 -fa on
How these were made
- Download
Qwen/Qwen3-Coder-30B-A3B-Instruct(original weights). - Convert to f16 GGUF with llama.cpp.
- Build an importance matrix over our calibration corpus.
- Quantize the ladder with
--imatrix. UD-Q4_K_XLadditionally pins the token-embedding and output tensors toQ8_0.
License
Original model by Qwen, released under the Apache 2.0 license. Full terms: Apache 2.0. Quantized by Atomic Chat.