What are the hardware requirements for Spark X2.5 4B Hadamard GSQ?

Spark X2.5 4B Hadamard GSQ requires approximately 4.5 GB of VRAM and supports a context window of 1024K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Spark X2.5 4B Hadamard GSQ and how large is it?

Spark X2.5 4B Hadamard GSQ is developed by F-Labs, a model with 4b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Spark X2.5 4B Hadamard GSQ?

Spark X2.5 4B Hadamard GSQ is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Spark X2.5 4B Hadamard GSQ by F-Labs — VRAM 4.5GB, 1024K context

Name: Spark X2.5 4B Hadamard GSQ
Author: F-Labs

Spark X2.5 4B Hadamard GSQ is an open-source language model by F-Labs. Features: 4b LLM, VRAM: 4.5GB, Context: 1024K, License: apache-2.0.

Arxiv:2208.07339 Arxiv:2210.17323 Arxiv:2306.00978 Arxiv:2306.03078 Arxiv:2307.09782 Arxiv:2312.09934 Arxiv:2404.00456 Arxiv:2405.16406 Base model:quantized:xhtoken/s... Base model:xhtoken/spark-x2.5-... Conversational Custom code Dv-ssq Dv ssq hadamard int8 int4 src Empirical-emergence En Groupwise-int4 Hadamard Int4 Int8 Kv-bss Low-rank Model-index Quantization Quarot Region:us Safetensors Selective-attention-preservati... Sharded Spark Spark2 5 Spinquant Src Svd Tensorflow Zh

Model Card on HF 🤗: https://huggingface.co/F-Labs/Spark-X2.5-4B-Hadamard-GSQ

Spark X2.5 4B Hadamard GSQ Parameters and Internals

LLM Name	Spark X2.5 4B Hadamard GSQ
Repository 🤗	https://huggingface.co/F-Labs/Spark-X2.5-4B-Hadamard-GSQ
Model Name	Spark-X2.5-4B-Hadamard-GSQ
Base Model(s)	XHToken/Spark-X2.5-4B XHToken/Spark-X2.5-4B
Model Size	4b
Required VRAM	4.5 GB
Updated	2026-09-12
Maintainer	F-Labs
Model Type	spark2_5
Model Files	1.4 GB: 1-of-5 1.0 GB: 2-of-5 1.0 GB: 3-of-5 1.0 GB: 4-of-5 0.1 GB: 5-of-5
Supported Languages	en zh
Model Architecture	Spark2_5ForCausalLM
License	apache-2.0
Context Length	1048576
Model Max Length	1048576
Transformers Version	4.57.1
Tokenizer Class	PreTrainedTokenizerFast
Beginning of Sentence Token	<｜start▁of▁sentence｜>
End of Sentence Token	<｜end▁of▁sentence｜>
Unk Token	<unk>
Vocabulary Size	131072

Rank the Spark X2.5 4B Hadamard GSQ capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Spark X2.5 4B Hadamard GSQ

Best Alternatives	Context / RAM	Downloads	Likes
Spark X2.5 4B	1024K / 8.3 GB	429	106
Spark X2.5 4B FP8	1024K / 4.4 GB	42	10
Spark X2.5 4B INT8	1024K / 4.5 GB	35	8
Spark X2.5 4B Base	1024K / 8.3 GB	16	5
Spark X2.5 4B MLX 8bit	1024K / 4.4 GB	858	3
Spark X2.5 4B MLX 4bit	1024K / 2.3 GB	617	2
Spark X2.5 4B Uncensored	1024K / 8.2 GB	238	0
...2.5 4B Abliterated Safetensors	1024K / 8.2 GB	47	2
...bliterated Uncensored MLX 4Bit	1024K / 8.2 GB	1028	1

Note: green Score (e.g. "73.2") means that the model is better than F-Labs/Spark-X2.5-4B-Hadamard-GSQ.