Spark X2.5 4B Heretic jp needs about 4.3 GB of memory at Q4_K_M (a 2.6 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 5 quants, from 2.6 GB (Q4_K_M) to 8.2 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.8 GB.
Spark X2.5 4B Heretic jp – Japanese‑adapted, tool‑calling LLM
Spark X2.5 4B Heretic jp is a 4.1 B parameter model adapted for Japanese with refusal attenuation and tool‑calling repair. It supports a native 1M‑token context and runs on llama.cpp with a custom fork. The model is released under Apache‑2.0.
Good at
- Tool‑calling in Japanese
- Long‑context reasoning up to 1M tokens
- Efficient inference on GPUs
Limits
- No vision support
- Requires XHToken/llama.cpp fork
- Large VRAM needed for higher quants
Best for
- Japanese coding assistants
- Agent workflows with tool use
- Long‑form Japanese text generation
Who it’s for. Developers building Japanese language agents or tools that need reliable tool‑calling. Researchers testing long‑context performance on a 4B parameter model. Users who prefer an open‑source model with an Apache‑2.0 license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 2.6 GB | 4.3 GB | Fits | Get file → | |
| Q5_K_M | 3.0 GB | 4.7 GB | Fits | Get file → | |
| Q6_K | 3.4 GB | 5.1 GB | Fits | Get file → | |
| Q8_0 | 4.4 GB | 6.1 GB | Fits | Get file → | |
| BF16 | 8.2 GB | 9.9 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (file size 2.6 GB, ~4.3 GB memory at 8k context). If 12 GB VRAM is available, Q5_K_M (2.98 GB, ~4.7 GB) offers higher quality. For 16 GB VRAM, Q6_K (3.38 GB, ~5.1 GB) provides a balance of speed and fidelity.
What GPU do you need for Spark X2.5 4B Heretic jp?
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 Q6_K (3.4 GB file, about 5.1 GB in use).
- 8 GB VRAM: the best quant that fits is Q8_0 (4.4 GB file, about 6.1 GB in use).
- 12 GB VRAM: the best quant that fits is BF16 (8.2 GB file, about 9.9 GB in use).
Which Spark X2.5 4B Heretic jp quant should you choose?
Q4_K_M is the usual starting point for Spark X2.5 4B Heretic jp: it balances file size and quality well. Here it is a 2.6 GB file.
The Q8_0 file is 4.4 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.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Spark X2.5 4B Heretic jp
How much VRAM does Spark X2.5 4B Heretic jp need?
About 4.3 GB at Q4_K_M with an 8K context window. That is the 2.6 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 Spark X2.5 4B Heretic jp on an 8 GB GPU?
Yes, with Q8_0: the 4.4 GB file needs about 6.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Spark X2.5 4B Heretic jp GGUF?
It depends on the quant: Q4_K_M is 2.6 GB, Q5_K_M is 3.0 GB, Q6_K is 3.4 GB, Q8_0 is 4.4 GB, BF16 is 8.2 GB.
What context length does Spark X2.5 4B Heretic jp support?
The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 1.2 GB at 8K and 4.8 GB at 32K.
Where can I download Spark X2.5 4B Heretic jp GGUF?
The files are on Hugging Face in the soyaakinohara/Spark-X2.5-4B-Heretic-jp-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 Spark X2.5 4B Heretic jp 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 Spark X2.5 4B Heretic jp on 8 GB VRAM?
Yes, the Q4_K_M quant uses about 4.3 GB of VRAM at 8k context, fitting within 8 GB.
Does this model support image inputs?
No, the model has no vision capability.
What software is required to run this model?
A Spark‑compatible llama.cpp fork (XHToken/llama.cpp) is required for inference.