TwIL LM3 needs about 3.0 GB of memory at Q4_K_M (a 1.9 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 19 quants, from 1.1 GB (IQ2_M) to 6.2 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 2.4 GB.
TwIL LM3 – a 3.08B formal‑logic reasoning model
TwIL LM3 is a 3.08B parameter model built from SmolLM3‑3B via LoRA SFT, checkpoint fusion, WiSE‑FT, and entropy‑weighted GRPO. It specializes in formal logic tasks such as FOL translation, entailment, semantic parsing, Lean formalisation, and proof critique. The model is not a general assistant and has no safety or preference tuning beyond SmolLM3.
Good at
- Formal logic reasoning
- Structured formal outputs
- Short answer generation
Limits
- No general chat support
- Instruction following regressed
- No vision capability
Best for
- Formal logic problem solving
- Semantic parsing tasks
- Lean proof critique
Who it’s for. Developers and researchers needing a lightweight formal‑logic model for local inference. The model is suitable for environments with limited VRAM, such as 4 GB or less, and for use with llama.cpp or compatible runtimes.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ2_M | 1.1 GB | 2.2 GB | Fits | Get file → | |
| Q2_K | 1.3 GB | 2.4 GB | Fits | Get file → | |
| IQ3_XXS | 1.3 GB | 2.4 GB | Fits | Get file → | |
| IQ3_XS | 1.4 GB | 2.5 GB | Fits | Get file → | |
| Q3_K_S | 1.4 GB | 2.5 GB | Fits | Get file → | |
| IQ3_M | 1.5 GB | 2.6 GB | Fits | Get file → | |
| Q3_K_M | 1.6 GB | 2.7 GB | Fits | Get file → | |
| Q3_K_L | 1.7 GB | 2.8 GB | Fits | Get file → | |
| IQ4_XS | 1.7 GB | 2.8 GB | Fits | Get file → | |
| Q4_0 | 1.8 GB | 2.9 GB | Fits | Get file → | |
| IQ4_NL | 1.8 GB | 2.9 GB | Fits | Get file → | |
| Q4_K_S | 1.8 GB | 2.9 GB | Fits | Get file → | |
| Q4_K_M | 1.9 GB | 3.0 GB | Fits | Get file → | |
| Q4_1 | 2.0 GB | 3.1 GB | Fits | Get file → | |
| Q5_K_S | 2.2 GB | 3.3 GB | Fits | Get file → | |
| Q5_K_M | 2.2 GB | 3.3 GB | Fits | Get file → | |
| Q6_K | 2.5 GB | 3.6 GB | Fits | Get file → | |
| Q8_0 | 3.3 GB | 4.4 GB | Fits | Get file → | |
| BF16 | 6.2 GB | 7.3 GB | Tight | Get file → |
For 4 GB VRAM use Q4_K_M (1.92 GB). With 8 GB VRAM, Q5_K_M (2.21 GB) or Q6_K (2.53 GB) provide higher quality. For 16 GB VRAM or more, Q8_0 (3.28 GB) or BF16 (6.16 GB) are available. Memory estimates are approximate.
What GPU do you need for TwIL LM3?
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 Q6_K (2.5 GB file, about 3.6 GB in use).
- 6 GB VRAM: the best quant that fits is Q8_0 (3.3 GB file, about 4.4 GB in use).
- 8 GB VRAM: the best quant that fits is BF16 (6.2 GB file, about 7.3 GB in use).
Which TwIL LM3 quant should you choose?
Q4_K_M is the usual starting point for TwIL LM3: it balances file size and quality well. Here it is a 1.9 GB file.
The Q8_0 file is 3.3 GB, about 1.7 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 IQ2_M at 1.1 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about TwIL LM3
How much VRAM does TwIL LM3 need?
About 3.0 GB at Q4_K_M with an 8K context window. That is the 1.9 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 TwIL LM3 on an 8 GB GPU?
Yes, with BF16: the 6.2 GB file needs about 7.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of TwIL LM3 GGUF?
It depends on the quant: IQ2_M is 1.1 GB, Q2_K is 1.3 GB, IQ3_XXS is 1.3 GB, IQ3_XS is 1.4 GB, Q3_K_S is 1.4 GB, IQ3_M is 1.5 GB and 13 more.
What context length does TwIL LM3 support?
The GGUF header lists a trained context length of 65,536 tokens (64K). The KV cache grows with the context you use: about 0.6 GB at 8K and 2.4 GB at 32K.
Where can I download TwIL LM3 GGUF?
The files are on Hugging Face in the NANI-Nithin/TwIL-LM3-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 TwIL LM3 use?
Hugging Face lists the license as other. 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 TwIL LM3 on 8 GB VRAM?
Yes, the Q4_K_M quant is 1.92 GB and can run on 8 GB VRAM. Higher quality quants such as Q5_K_M or Q6_K also fit within 8 GB.
Does TwIL LM3 support vision tasks?
No, TwIL LM3 does not have vision capability.
What is the context length of TwIL LM3?
The model supports a 65,536‑token context length.