K2 Horizon 32B Heretic needs about 23.7 GB of memory at Q4_K_M (a 21.1 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 11 quants, from 13.3 GB (Q2_K) to 37.0 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.
K2 Horizon 32B Heretic – 34.8B parameter model with refusal attenuation
K2 Horizon 32B Heretic is a 34.8B parameter model based on the K2-Horizon architecture. It includes a refusal‑attenuated derivative of the IFM/K2-Horizon-32B model and provides coding, reasoning, and tools abilities. Full‑precision weights require ~70 GB of memory.
Good at
- Coding
- Reasoning
- Tool use
Limits
- Full‑precision needs ~70 GB VRAM
- No vision support
- Apache‑2.0 license
Best for
- Local LLM inference
- Code generation
- Tool‑integrated workflows
Who it’s for. Developers and researchers who need a large, locally‑hosted model with refusal attenuation and coding capabilities. The model is suitable for users with sufficient VRAM or those who can use GGUF quantizations.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 13.3 GB | 15.9 GB | Too big | Get file → | |
| Q3_K_S | 15.5 GB | 18.1 GB | Too big | Get file → | |
| Q3_K_M | 17.1 GB | 19.7 GB | Too big | Get file → | |
| Q3_K_L | 18.5 GB | 21.1 GB | Too big | Get file → | |
| IQ4_XS | 19.1 GB | 21.7 GB | Too big | Get file → | |
| Q4_K_S | 20.0 GB | 22.7 GB | Too big | Get file → | |
| Q4_K_M | 21.1 GB | 23.7 GB | Too big | Get file → | |
| Q5_K_S | 24.1 GB | 26.7 GB | Too big | Get file → | |
| Q5_K_M | 24.7 GB | 27.3 GB | Too big | Get file → | |
| Q6_K | 28.5 GB | 31.2 GB | Too big | Get file → | |
| Q8_0 | 37.0 GB | 39.6 GB | Too big | Get file → |
For 8 GB VRAM use Q4_K_M (file size 21.08 GB, ~23.7 GB memory at 8k context). For 16 GB VRAM use Q4_K_S (20.05 GB, ~22.7 GB). For larger VRAM, Q8_0 (36.97 GB, ~39.6 GB) offers full precision. Estimates are approximate.
What GPU do you need for K2 Horizon 32B Heretic?
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: the best quant that fits is Q2_K (13.3 GB file, about 15.9 GB in use).
- 24 GB VRAM: the best quant that fits is Q4_K_M (21.1 GB file, about 23.7 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K (28.5 GB file, about 31.2 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (37.0 GB file, about 39.6 GB in use).
Which K2 Horizon 32B Heretic quant should you choose?
Q4_K_M is the usual starting point for K2 Horizon 32B Heretic: it balances file size and quality well. Here it is a 21.1 GB file.
The Q8_0 file is 37.0 GB, about 1.8 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 Q2_K at 13.3 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 K2 Horizon 32B Heretic
How much VRAM does K2 Horizon 32B Heretic need?
About 23.7 GB at Q4_K_M with an 8K context window. That is the 21.1 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 K2 Horizon 32B Heretic on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 15.9 GB, so you would need a larger GPU or CPU offloading.
What is the file size of K2 Horizon 32B Heretic GGUF?
It depends on the quant: Q2_K is 13.3 GB, Q3_K_S is 15.5 GB, Q3_K_M is 17.1 GB, Q3_K_L is 18.5 GB, IQ4_XS is 19.1 GB, Q4_K_S is 20.0 GB and 5 more.
What context length does K2 Horizon 32B Heretic support?
The GGUF header lists a trained context length of 524,288 tokens (512K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.
Where can I download K2 Horizon 32B Heretic GGUF?
The files are on Hugging Face in the mradermacher/K2-Horizon-32B-heretic-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 K2 Horizon 32B Heretic 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 this model on 8 GB VRAM?
Use the Q4_K_M GGUF quant; it requires about 23.7 GB of memory at 8k context, which is beyond 8 GB VRAM, so it will need CPU fallback or a larger GPU.
Does the model support vision?
No, the model has no vision capabilities.
What license applies?
The model is released under the Apache‑2.0 license.