Spark-X2.6 4B VRAM Requirements

Spark-X2.6 4B at a glance

Made by

iFlytek, China

Released

October 19, 2026

Parameters

4.11B

Context window

1M tokens

License

Apache 2.0, commercial use allowed

On Hugging Face

2 likes, 0 downloads last month (opens in a new tab)

On ModelScope

XHToken/Spark-X2.6-4B (opens in a new tab)

Runs with

vLLM, SGLang, llama.cpp (Ollama, LM Studio)

vLLM needs the extra setup described on the model card (model card (opens in a new tab))

VRAM by precision and context length

Weights plus the KV cache of one request (BF16 cache), in GiB. Add 1 to 3 GiB for the inference engine itself.

Spark-X2.6 4B GGUF files

Exact size of each quantization in XHToken/Spark-X2.6-4B-GGUF (opens in a new tab), and the smallest GPU that runs it with llama.cpp, Ollama or LM Studio and an 8K-token context.

Read every day from Hugging Face. Split files are added together; vision projectors are left out.

How many GPUs to run Spark-X2.6 4B

Fewest GPUs that fit the weights plus one 8K-token request. Data-center GPUs use vLLM (90% of memory usable), the others llama.cpp. Several GPUs split the model with tensor or pipeline parallelism.

8-bit: FP8 with vLLM (INT8 on GPUs without FP8), Q8_0 with llama.cpp. 4-bit: AWQ / GPTQ with vLLM, Q4_K_M with llama.cpp, both counted at the Q4_K_M size.

Cheapest way to run Spark-X2.6 4B today

For each GPU: the fewest cards that fit Spark-X2.6 4B with one 8K-token request, the most faithful weight format at that count, and the cheapest on-demand price (2026-10-20).

On-demand prices from AWS, Google Cloud, Azure, RunPod, Vast.ai, Oracle Cloud, OVHcloud, Scaleway, Verda, Akash and UpCloud, checked every hour. Serving many users needs more KV cache, so more memory: size it in the calculator. Some links are affiliate links: StudioTV may earn a commission, at no extra cost to you.

AWS, Google Cloud and UpCloud prices: Spare Cores (opens in a new tab), CC BY-SA 4.0 (opens in a new tab), shared under the same license.

KV cache: how memory grows with context

Architecture: 36 layers · 9 full attention (4 KV × 256) · 27 sliding-window layers (512 tokens, 4 KV × 256). Its KV cache is 3.8× smaller than if every layer used full attention, because most layers only keep a short window of recent tokens.

Spark-X2.6 4B VRAM badge

For a model card or a README: the 4-bit size and the smallest GPU it fits on, linked to this page. Data as JSON: /api/models/spark-x2-6-4b.json.

[![Spark-X2.6 4B VRAM](https://studiotvai.com/badge/spark-x2-6-4b.svg)](https://studiotvai.com/vram-requirements/spark-x2-6-4b)

Other models

Same family (sliding-window attention): Gemma 2 9B · Gemma 3 4B · Gemma 4 12B · Gemma 4 31B · Gemma 4 26B-A4B · gpt-oss 20B · gpt-oss 120B · Llama 4 Scout 17B-16E · Llama 4 Maverick 17B-128E · Kolibri-1 · Nemotron Labs Dory 1.0 4B · Step 3.5 Flash · Muse Glimmer 30B · Humanizer · Mellum2.1 12B-A2.5B · Spark-X2.5 4B · Ministral 8B · Gemma 4 E4B · Gemma 4 E2B · Unlimited-OCR

Estimates, not guarantees: computed from the official config.json with the same engine as the LLM VRAM Calculator. Real usage depends on your engine version and settings.