avnigashi/spark-x25-4b-chat

Spark-X2.5-4B Chat

An interactive demo for XHToken/Spark-X2.5-4B, a compact general-purpose language model with strong reasoning, coding, multilingual, and agent capabilities.

The app runs the model directly in a Hugging Face ZeroGPU Space. Choose whether to expose the model's reasoning trace, adjust generation controls, and try the included prompts.

Notes

• —The model is loaded using the official Transformers configuration and chat template supplied by its repository.
• —The context window is limited to 16,384 tokens in this demo to keep interactive requests practical. The model itself supports a much larger native context window.
• —Inference uses bfloat16 on the temporary GPU allocated by ZeroGPU.

License

This demo code is Apache-2.0. The underlying model is released under its Apache-2.0 license.