Laguna-S-2.1 needs about 75.2 GB of memory at Q4_K_M (a 73.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 23 quants, from 33.8 GB (IQ1_S) to 235.2 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 6.4 GB.
Laguna‑S‑2.1: 118B Mixture‑of‑Experts model for long‑horizon coding
Laguna‑S‑2.1 is a 118B total parameter Mixture‑of‑Experts model with 8B activated parameters per token. It features 48 layers, 256 routed experts plus one shared expert, and a 1,048,576‑token context window. The model is released under the fully permissive OpenMDW‑1.1 license.
Good at
- Long‑horizon coding
- Agentic reasoning
- Large context handling
Limits
- Requires large VRAM for full precision
- No vision support
- Limited to text‑to‑text modalities
Best for
- Complex code generation
- Multi‑step reasoning tasks
- Large‑scale text analysis
Who it’s for. Developers and researchers needing a high‑capacity language model for coding and reasoning tasks. The model can be used in commercial products under the OpenMDW‑1.1 license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 33.8 GB | 35.9 GB | Too big | Get file → | |
| IQ1_M | 35.6 GB | 37.8 GB | Too big | Get file → | |
| IQ2_XXS | 37.2 GB | 39.3 GB | Too big | Get file → | |
| IQ2_M | 37.3 GB | 39.4 GB | Too big | Get file → | |
| Q2_K_XL | 39.7 GB | 41.8 GB | Too big | Get file → | |
| IQ3_XXS | 44.3 GB | 46.4 GB | Too big | Get file → | |
| IQ3_S | 48.4 GB | 50.5 GB | Too big | Get file → | |
| Q3_K_M | 54.0 GB (3 parts) | 56.1 GB | Too big | Get file → | |
| Q3_K_XL | 54.1 GB (3 parts) | 56.2 GB | Too big | Get file → | |
| IQ4_XS | 57.6 GB (3 parts) | 59.7 GB | Too big | Get file → | |
| IQ4_NL | 58.7 GB (3 parts) | 60.9 GB | Too big | Get file → | |
| Q4_K_S | 68.6 GB (3 parts) | 70.7 GB | Too big | Get file → | |
| MXFP4_MOE | 71.1 GB (3 parts) | 73.2 GB | Too big | Get file → | |
| Q4_K_M | 73.1 GB (3 parts) | 75.2 GB | Too big | Get file → | |
| Q4_K_XL | 73.4 GB (3 parts) | 75.5 GB | Too big | Get file → | |
| Q5_K_S | 82.7 GB (3 parts) | 84.8 GB | Too big | Get file → | |
| Q5_K_M | 87.9 GB (3 parts) | 90.0 GB | Too big | Get file → | |
| Q5_K_XL | 88.1 GB (3 parts) | 90.2 GB | Too big | Get file → | |
| Q6_K | 97.9 GB (3 parts) | 100.0 GB | Too big | Get file → | |
| Q6_K_XL | 107.1 GB (4 parts) | 109.2 GB | Too big | Get file → | |
| Q8_0 | 125.0 GB (4 parts) | 127.1 GB | Too big | Get file → | |
| Q8_K_XL | 128.1 GB (4 parts) | 130.2 GB | Too big | Get file → | |
| BF16 | 235.2 GB (5 parts) | 237.3 GB | Too big | Get file → |
For 8 GB VRAM, use the IQ1_S or IQ1_M quant, which require about 36–38 GB of VRAM at 8 k context. For 16 GB VRAM, Q4_K_M or Q4_K_XL are suitable, needing roughly 75 GB of VRAM. Larger VRAM allows higher‑quality quants such as Q5_K_XL or Q6_K_XL.
What GPU do you need for Laguna-S-2.1?
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: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: the best quant that fits is IQ3_XXS (44.3 GB file, about 46.4 GB in use).
Which Laguna-S-2.1 quant should you choose?
Q4_K_M is the usual starting point for Laguna-S-2.1: it balances file size and quality well. Here it is a 73.1 GB file.
The Q8_0 file is 125.0 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 IQ1_S at 33.8 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 Laguna-S-2.1
How much VRAM does Laguna-S-2.1 need?
About 75.2 GB at Q4_K_M with an 8K context window. That is the 73.1 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 Laguna-S-2.1 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 35.9 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Laguna-S-2.1 GGUF?
It depends on the quant: IQ1_S is 33.8 GB, IQ1_M is 35.6 GB, IQ2_XXS is 37.2 GB, IQ2_M is 37.3 GB, Q2_K_XL is 39.7 GB, IQ3_XXS is 44.3 GB and 17 more.
What context length does Laguna-S-2.1 support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.6 GB at 8K and 6.4 GB at 32K.
Where can I download Laguna-S-2.1 GGUF?
The files are on Hugging Face in the unsloth/Laguna-S-2.1-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 Laguna-S-2.1 use?
Hugging Face lists the license as openmdw-1.1. 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 Laguna‑S‑2.1 on 8 GB VRAM?
No, the smallest quant requires about 36 GB of VRAM at 8 k context. 8 GB is insufficient for any variant.
Does Laguna‑S‑2.1 support image inputs?
No, the model is text‑to‑text only and has no vision capabilities.