Avrik Lite 4B needs about 7.2 GB of memory at Q4_K_M (a 5.3 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 16 quants, from 0.0 GB to 6.2 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 5.6 GB.
Avrik Lite 4B – a 7.52B‑parameter decision model in GGUF format
Avrik Lite 4B is a decision model built on Gemma 4 E4B‑it. It returns calibrated probability distributions for typed questions without generating text. The model is fast on ordinary CPUs and designed for high‑throughput decision tasks.
Good at
- Calibrated probability decisions
- Fast inference on CPU
- Low memory footprint
Limits
- No text generation
- Requires GGUF runtime
- Limited to decision tasks
Best for
- Routing decisions
- Policy approvals
- Moderation triage
Who it’s for. Developers building decision‑making systems that need calibrated probabilities and low latency. Suitable for edge deployments with limited GPU memory.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 1.9 GB | Fits | Get file → | |
| Q2_K | 4.4 GB | 6.3 GB | Fits | Get file → | |
| Q3_K_S | 4.7 GB | 6.6 GB | Fits | Get file → | |
| IQ3_S | 4.7 GB | 6.6 GB | Fits | Get file → | |
| IQ3_M | 4.7 GB | 6.6 GB | Fits | Get file → | |
| Q3_K_M | 4.9 GB | 6.8 GB | Fits | Get file → | |
| Q3_K_L | 5.0 GB | 6.9 GB | Fits | Get file → | |
| IQ4_XS | 5.1 GB | 7.0 GB | Fits | Get file → | |
| IQ4_NL | 5.2 GB | 7.1 GB | Fits | Get file → | |
| Q4_0 | 5.2 GB | 7.1 GB | Fits | Get file → | |
| Q4_K_S | 5.2 GB | 7.1 GB | Fits | Get file → | |
| Q4_K_M | 5.3 GB | 7.2 GB | Tight | Get file → | |
| Q4_1 | 5.4 GB | 7.3 GB | Tight | Get file → | |
| Q5_K_S | 5.7 GB | 7.6 GB | Tight | Get file → | |
| Q5_K_M | 5.8 GB | 7.7 GB | Tight | Get file → | |
| Q6_K | 6.2 GB | 8.1 GB | Too big | Get file → |
For 8 GB VRAM use Q4_0 (est. 7.1 GB memory at 8 k context) or Q4_K_S. For 6 GB VRAM use Q3_K_M (est. 6.6 GB). For 4 GB VRAM use Q2_K (est. 6.3 GB). File sizes are 5.19 GB, 4.85 GB, and 4.4 GB respectively.
What GPU do you need for Avrik Lite 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 the main file (0.0 GB file, about 1.9 GB in use).
- 6 GB VRAM: the best quant that fits is the main file (0.0 GB file, about 1.9 GB in use).
- 8 GB VRAM: the best quant that fits is Q5_K_M (5.8 GB file, about 7.7 GB in use).
- 12 GB VRAM: the best quant that fits is Q6_K (6.2 GB file, about 8.1 GB in use).
Which Avrik Lite 4B quant should you choose?
Q4_K_M is the usual starting point for Avrik Lite 4B: it balances file size and quality well. Here it is a 5.3 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 Avrik Lite 4B
How much VRAM does Avrik Lite 4B need?
About 7.2 GB at Q4_K_M with an 8K context window. That is the 5.3 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 Avrik Lite 4B on an 8 GB GPU?
Yes, with Q5_K_M: the 5.8 GB file needs about 7.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Avrik Lite 4B GGUF?
It depends on the quant: file is 0.0 GB, Q2_K is 4.4 GB, Q3_K_S is 4.7 GB, IQ3_S is 4.7 GB, IQ3_M is 4.7 GB, Q3_K_M is 4.9 GB and 10 more.
What context length does Avrik Lite 4B support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 1.4 GB at 8K and 5.6 GB at 32K.
Where can I download Avrik Lite 4B GGUF?
The files are on Hugging Face in the mradermacher/Avrik-Lite-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 Avrik Lite 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 Avrik Lite 4B on 8 GB VRAM?
Yes, with Q4_0 or Q4_K_S quants the model uses about 7.1 GB of VRAM at 8 k context.
Does this model support text generation?
No, it outputs calibrated probabilities only; use avrik-server for readout.
What license does it use?
Apache-2.0