Haru Qwen3 4b needs about 4.2 GB of memory at Q4_K_M (a 2.5 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 25 quants, from 0.0 GB to 3.3 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 4.8 GB.
Haru Qwen3 4b – lightweight Korean character chatbot
Haru Qwen3 4b is a 4 billion‑parameter model fine‑tuned for the HaruPet app. It speaks in short, supportive sentences in Korean, ending each reply with a mood gauge. The model is based on Qwen3‑4B‑Instruct‑2507 and uses LoRA training.
Good at
- Short Korean roleplay dialogue
- Character‑specific tone
- Mood gauge output
Limits
- May produce missing words or unrelated answers
- No real‑time data access
- Not suitable for medical or legal advice
Best for
- Chatbot in mobile pet apps
- Short character interactions
- Low‑resource inference
Who it’s for. Developers building lightweight Korean character chatbots or mobile pet apps. The model is ideal for environments with limited VRAM and for users who need a consistent, short‑form conversational style.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 1.7 GB | Fits | Get file → | |
| IQ1_S | 1.1 GB | 2.8 GB | Fits | Get file → | |
| IQ1_M | 1.1 GB | 2.8 GB | Fits | Get file → | |
| IQ2_XXS | 1.2 GB | 3.0 GB | Fits | Get file → | |
| IQ2_XS | 1.4 GB | 3.1 GB | Fits | Get file → | |
| IQ2_S | 1.4 GB | 3.1 GB | Fits | Get file → | |
| IQ2_M | 1.5 GB | 3.2 GB | Fits | Get file → | |
| Q2_K_S | 1.6 GB | 3.3 GB | Fits | Get file → | |
| Q2_K | 1.7 GB | 3.4 GB | Fits | Get file → | |
| IQ3_XXS | 1.7 GB | 3.4 GB | Fits | Get file → | |
| IQ3_XS | 1.8 GB | 3.5 GB | Fits | Get file → | |
| Q3_K_S | 1.9 GB | 3.6 GB | Fits | Get file → | |
| IQ3_S | 1.9 GB | 3.6 GB | Fits | Get file → | |
| IQ3_M | 2.0 GB | 3.7 GB | Fits | Get file → | |
| Q3_K_M | 2.1 GB | 3.8 GB | Fits | Get file → | |
| Q3_K_L | 2.2 GB | 3.9 GB | Fits | Get file → | |
| IQ4_XS | 2.3 GB | 4.0 GB | Fits | Get file → | |
| Q4_0 | 2.4 GB | 4.1 GB | Fits | Get file → | |
| IQ4_NL | 2.4 GB | 4.1 GB | Fits | Get file → | |
| Q4_K_S | 2.4 GB | 4.1 GB | Fits | Get file → | |
| Q4_K_M | 2.5 GB | 4.2 GB | Fits | Get file → | |
| Q4_1 | 2.6 GB | 4.3 GB | Fits | Get file → | |
| Q5_K_S | 2.8 GB | 4.5 GB | Fits | Get file → | |
| Q5_K_M | 2.9 GB | 4.6 GB | Fits | Get file → | |
| Q6_K | 3.3 GB | 5.0 GB | Fits | Get file → |
For 8 GB VRAM, use Q4_K_M (file size 2.5 GB, estimated 4.2 GB memory). For 6 GB VRAM, Q3_K_M (2.08 GB, 3.8 GB memory). If memory is very tight, IQ1_S (1.06 GB, 2.8 GB memory) is the smallest option.
What GPU do you need for Haru Qwen3 4b?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: the best quant that fits is IQ4_XS (2.3 GB file, about 4.0 GB in use).
- 6 GB VRAM: the best quant that fits is Q6_K (3.3 GB file, about 5.0 GB in use).
Which Haru Qwen3 4b quant should you choose?
Q4_K_M is the usual starting point for Haru Qwen3 4b: it balances file size and quality well. Here it is a 2.5 GB file.
If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Haru Qwen3 4b
How much VRAM does Haru Qwen3 4b need?
About 4.2 GB at Q4_K_M with an 8K context window. That is the 2.5 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run Haru Qwen3 4b on an 8 GB GPU?
Yes, with Q6_K: the 3.3 GB file needs about 5.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Haru Qwen3 4b GGUF?
It depends on the quant: file is 0.0 GB, IQ1_S is 1.1 GB, IQ1_M is 1.1 GB, IQ2_XXS is 1.2 GB, IQ2_XS is 1.4 GB, IQ2_S is 1.4 GB and 19 more.
What context length does Haru Qwen3 4b support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.2 GB at 8K and 4.8 GB at 32K.
Where can I download Haru Qwen3 4b GGUF?
The files are on Hugging Face in the mradermacher/haru-qwen3-4b-i1-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Haru Qwen3 4b use?
Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run Haru Qwen3 4b on 8 GB VRAM?
Yes, using a quant such as Q4_K_M or Q3_K_M will fit within 8 GB VRAM with an estimated memory usage of 4.2 GB or 3.8 GB respectively.
Does the model support vision?
No, Haru Qwen3 4b does not have vision capabilities.
What license applies to this model?
The model is released under the Apache‑2.0 license.