Qwen3-4B-GGUF
See our collection for all versions of Qwen3 including GGUF, 4-bit & 16-bit formats. Learn to run Qwen3 correctly - Read our Guide. Unsloth Dynamic 2.0 achieves superior accuracy & outperforms other leading quants. ✨ Run & Fine-tune Qwen3 with Unsloth!
Params
—
Context
40,960
Downloads 30d
481 K
Likes
244
Download history
daily snapshots · 37 days
▲ 196 K in the last 30 days (68.7%)
512 K310 K
Aug 22Sep 1Sep 11Sep 20
512 K185 K
Aug 15Aug 27Sep 8Sep 20
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| Qwen3-4B-Q2_K.gguf | Q2_K | 1.7 GB | 2.3 GB | ✅ Runs comfortably |
| Qwen3-4B-Q2_K_L.gguf | Q2_K_L | 1.7 GB | 2.3 GB | ✅ Runs comfortably |
| Qwen3-4B-Q3_K_S.gguf | Q3_K_S | 1.9 GB | 2.6 GB | ✅ Runs comfortably |
| Qwen3-4B-Q3_K_M.gguf | Q3_K_M | 2.1 GB | 2.8 GB | ✅ Runs comfortably |
| Qwen3-4B-IQ4_XS.gguf | IQ4_XS | 2.3 GB | 3.0 GB | ✅ Runs comfortably |
| Qwen3-4B-IQ4_NL.gguf | IQ4_NL | 2.4 GB | 3.1 GB | ✅ Runs comfortably |
| Qwen3-4B-Q4_0.gguf | Q4_0 | 2.4 GB | 3.1 GB | ✅ Runs comfortably |
| Qwen3-4B-BF16.gguf | GGUF | 8.1 GB | 9.4 GB | ✅ Runs comfortably |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ ollama run qwen3-4b-gguf-unsloth # pin the quantization explicitly $ ollama run qwen3-4b-gguf-unsloth-q4_0
$ huggingface-cli download unsloth/Qwen3-4B-GGUF-GGUF \
Qwen3-4B-Q4_0.gguf --local-dir .
$ llama-cli -m Qwen3-4B-Q4_0.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/qwen3-4b-gguf-unsloth
{
"hf_id": "unsloth/Qwen3-4B-GGUF",
"params_b": null,
"context_length": 40960,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 481224,
"vram_estimates": [
{ "quant": "GGUF", "gb": 9.4 },
{ "quant": "IQ4_NL", "gb": 3.1 }
],
"updated_at": "2026-08-15T01:00:30Z"
}
Specifications
- Architecture
- Qwen3ForCausalLM
- Context length
- 40,960
- Vocabulary
- 151,936
- Layers / heads
- 36 / 32
- Licence
- apache-2.0
- First seen on the Hub
- 2025-04-28
- Base model
- Qwen3-4B
- Training datasets
- undisclosed
- Added to our catalog
- 2026-08-15
Compare with any text-generation model