Ahma 3B Instruct needs about 6.0 GB of memory at Q4_K_M (a 2.8 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 22 quants, from 2.1 GB (IQ3_S) to 3.9 GB (Q8_0).
Ahma 3B Instruct – Finnish instruction‑tuned Llama model
Ahma 3B Instruct is a 3.63 billion‑parameter decoder‑only transformer based on Meta Llama v1. It is tuned to follow instructions in Finnish. The model is released under the Apache‑2.0 license.
Good at
- Instruction following in Finnish
- Chat‑style responses
- Text generation in Finnish
Limits
- No vision support
- Context limited to 2048 tokens
- Requires at least 5.4 GB VRAM for Q2_K at 8k context
Best for
- Local Finnish chatbots
- Finnish content creation
- Instruction‑driven data analysis
Who it’s for. Developers building Finnish language applications who need a lightweight, instruction‑tuned model. Researchers working on Finnish NLP can use it for fine‑tuning or evaluation.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ3_S | 2.1 GB | 3.3 GB | Fits | Get file → | |
| IQ3_XS | 2.1 GB | 3.3 GB | Fits | Get file → | |
| Q2_K | 2.1 GB | 3.3 GB | Fits | Get file → | |
| Q3_K_S | 2.1 GB | 3.3 GB | Fits | Get file → | |
| Q4_0 | 2.1 GB | 3.3 GB | Fits | Get file → | |
| IQ4_NL | 2.2 GB | 3.3 GB | Fits | Get file → | |
| IQ4_XS | 2.2 GB | 3.3 GB | Fits | Get file → | |
| IQ3_M | 2.2 GB | 3.4 GB | Fits | Get file → | |
| Q3_K | 2.3 GB | 3.5 GB | Fits | Get file → | |
| Q3_K_M | 2.3 GB | 3.5 GB | Fits | Get file → | |
| Q4_1 | 2.4 GB | 3.5 GB | Fits | Get file → | |
| Q3_K_L | 2.4 GB | 3.6 GB | Fits | Get file → | |
| Q5_0 | 2.6 GB | 3.8 GB | Fits | Get file → | |
| Q4_K_S | 2.6 GB | 3.8 GB | Fits | Get file → | |
| Q4_K | 2.8 GB | 3.9 GB | Fits | Get file → | |
| Q4_K_M | 2.8 GB | 3.9 GB | Fits | Get file → | |
| Q5_1 | 2.8 GB | 4.0 GB | Fits | Get file → | |
| Q5_K_S | 2.8 GB | 4.0 GB | Fits | Get file → | |
| Q5_K | 2.9 GB | 4.1 GB | Fits | Get file → | |
| Q5_K_M | 2.9 GB | 4.1 GB | Fits | Get file → | |
| Q6_K | 3.9 GB | 5.0 GB | Fits | Get file → | |
| Q8_0 | 3.9 GB | 5.0 GB | Fits | Get file → |
For 8 GB VRAM, use Q2_K, IQ3_S, or IQ3_XS (2.15 GB file, ~5.4 GB VRAM). For 12 GB VRAM, Q4_K_M (2.76 GB file, ~6 GB VRAM) or Q5_K_M (2.95 GB file, ~6.2 GB VRAM). Larger VRAM allows Q6_K or Q8_0 (3.86 GB file, ~7.1 GB VRAM).
What GPU do you need for Ahma 3B Instruct?
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: no quant fits at an 8K context.
- 6 GB VRAM: the best quant that fits is Q5_1 (2.8 GB file, about 6.0 GB in use).
- 8 GB VRAM: the best quant that fits is Q6_K (3.9 GB file, about 7.1 GB in use).
Which Ahma 3B Instruct quant should you choose?
Q4_K_M is the usual starting point for Ahma 3B Instruct: it balances file size and quality well. Here it is a 2.8 GB file.
The Q8_0 file is 3.9 GB, about 1.4 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is IQ3_S at 2.1 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about Ahma 3B Instruct
How much VRAM does Ahma 3B Instruct need?
About 6.0 GB at Q4_K_M with an 8K context window. That is the 2.8 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 Ahma 3B Instruct on an 8 GB GPU?
Yes, with Q6_K: the 3.9 GB file needs about 7.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Ahma 3B Instruct GGUF?
It depends on the quant: IQ3_S is 2.1 GB, IQ3_XS is 2.1 GB, Q2_K is 2.1 GB, Q3_K_S is 2.1 GB, Q4_0 is 2.1 GB, IQ4_NL is 2.2 GB and 16 more.
What context length does Ahma 3B Instruct support?
The GGUF header lists a trained context length of 2,048 tokens (2K). The KV cache grows with the context you use: about 2.7 GB at 8K and 0.7 GB at 2K.
Where can I download Ahma 3B Instruct GGUF?
The files are on Hugging Face in the RichardErkhov/Finnish-NLP_-_Ahma-3B-Instruct-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 Ahma 3B Instruct 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 Ahma 3B Instruct on 8 GB VRAM?
Yes, using Q2_K, IQ3_S, or IQ3_XS quantizations which require about 5.4 GB VRAM at 8k context.
Does the model support English?
The model is tuned for Finnish; English support is not guaranteed.
Is the model free to use?
Yes, it is licensed under Apache‑2.0, allowing free use and modification.