---
base_model:
- XHToken/Spark-X2.5-4B
license: apache-2.0

library_name: gguf
pipeline_tag: text-generation
tags:
- gguf
- llama.cpp
- spark-x2.5
- long-context
- 1m-context
---

> [!IMPORTANT]
> **Compatibility:** These GGUF files require `llama.cpp` **b10828** or later, which includes official support for the Spark-X2.5 (`spark2_5`) architecture. Applications with a bundled runtime must use an equivalent or newer build. [llama.cpp support](https://github.com/ggml-org/llama.cpp/pull/27868)

# Spark-X2.5-4B GGUF

GGUF quantizations of [XHToken/Spark-X2.5-4B](https://huggingface.co/XHToken/Spark-X2.5-4B), a 4B general-purpose language model for reasoning, coding, tool use, and agentic workflows. Native context: **1,048,576 tokens (1M)**.

## Benchmarks

![Spark-X2.5-4B benchmark results](assets/benchmark.png)

*Benchmark results reported by XHToken for Spark-X2.5-4B in thinking mode.*

## GGUF files

| Quantization | File | Size |
| --- | --- | ---: |
| Q4_0 | [Spark-X2.5-4B-Q4_0.gguf](Spark-X2.5-4B-Q4_0.gguf) | 2.41 GB |
| Q4_K_M | [Spark-X2.5-4B-Q4_K_M.gguf](Spark-X2.5-4B-Q4_K_M.gguf) | 2.60 GB |
| Q5_K_M | [Spark-X2.5-4B-Q5_K_M.gguf](Spark-X2.5-4B-Q5_K_M.gguf) | 2.98 GB |
| Q6_K | [Spark-X2.5-4B-Q6_K.gguf](Spark-X2.5-4B-Q6_K.gguf) | 3.38 GB |
| Q8_0 | [Spark-X2.5-4B-Q8_0.gguf](Spark-X2.5-4B-Q8_0.gguf) | 4.38 GB |

Includes the upstream [chat_template.jinja](chat_template.jinja). Checksums: [SHA256SUMS.txt](SHA256SUMS.txt).

## Usage

```bash
llama-cli -m Spark-X2.5-4B-Q4_K_M.gguf -c 131072 -cnv
```

## Source

- Model: [XHToken/Spark-X2.5-4B](https://huggingface.co/XHToken/Spark-X2.5-4B)
- Revision: `ea14618d20e76b5b093d3ee20a5b9d733bb12410`
- License: [Apache-2.0](LICENSE)