selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

XHToken logo

Spark-X2.6-4B-Base

XHToken·Base model·released 19 Oct 2026

Apache 2.04.1B1M ctx ≈4.1 GB
Your GPU
ollama run hf.co/mradermacher/Spark-X2.6-4B-Base-i1-GGUF:Q4_K_M
All 5 runners ↓

A little over a month has passed since Spark-X2.5. In the rapidly evolving world of AI agents, that is enough time to rethink, refine, and move the frontier forward.

Notify me when

How much VRAM does Spark-X2.6-4B-Base need?

At Q4_K_M Spark-X2.6-4B-Base needs about 4.1 GB with an 8K context (4.1 GB at Q4, 9.7 GB at full precision). Pick your GPU to see what fits and how fast it runs.

--vram
any
--gpus
--quant

Needs: ≈4.2 GB · Q4_K_M · 8K context — Pick a GPU to check the fit.

Change quant, context or KV cache to compare
Advanced: context 8K · KV cache 16-bit
Context
KV cache

Smallest GPU for each quant and context (KV cache 16-bit)

Quantisation8K16K32K64K128K1M
Q3_K_M
Q4_K_M
Q5_K_M
Q6_K
Q8_0
FP16

≤16 GB card24–32 GB48–96 GBmore

QuantisationQualityWeightsKV cacheTotal
Q2_KHeavy lossSmallest, noticeably worse qualityHeavy loss1.7 GB file1.1 GB≈3.4 GB
Q3_K_MNoticeable lossSmall, some quality lossNoticeable loss2.0 GB file1.1 GB≈3.8 GB
Q4_K_MrecommendedSmall lossRecommended default for local useSmall loss2.4 GB file1.1 GB≈4.2 GB
Q5_K_MNear-losslessNear-lossless for most tasksNear-lossless2.8 GB file1.1 GB≈4.5 GB
Q6_KNear-losslessPractically losslessNear-lossless3.1 GB file1.1 GB≈4.9 GB
Q8_0LosslessLossless in practice, twice the size of Q4Lossless4.1 GB est.1.1 GB≈5.9 GB
FP16 / BF16LosslessFull precision, as releasedLossless7.7 GB est.1.1 GB≈9.7 GB

Estimates for 8K tokens of context: weights + KV cache + runtime overhead. Speed assumes memory-bound decoding at your GPU’s bandwidth. Pick a GPU above and the fit column checks every row against it. Weights marked “file” use the real GGUF download size; “est.” rows are computed from the parameter count.

How to pick a quant

Take the biggest quant that fits with room left for context. Q4_K_M is the usual sweet spot. Below Q4 quality drops fast, above Q6 you mostly pay memory for nothing. Big models (70B+, large MoE) tolerate low quants better than small ones.

  • LosslessF16, Q8_0
  • Near-losslessQ6_K, Q5_K_M
  • Small lossQ5_K_S, Q4_K_M, IQ4_NL, IQ4_XS, Q4_K_S
  • Noticeable lossQ4_0, Q3_K_L, Q3_K_M, IQ3_M
  • Heavy lossIQ3_XXS, Q3_K_S, IQ2_M, Q2_K

Typical quality loss, not measured per model. Tiers come from KL divergence against FP16 on Llama 3 8B: llama.cpp perplexity README: LLaMA 3 8B scoreboard (KLD and ΔPPL vs FP16, Wikitext-2); llama.cpp quantize tool: ΔPPL per quant type @ Llama-3-8B

How much context fits?

Model quant
KV cache

At 8K context, Q4_K_M needs ≈4.2 GB(2.4 GB weights + 1.1 GB KV cache + 0.6 GB overhead). Pick a GPU above to see how much context fits.

0501001502001K2K4K8K16K32K64K128K256K512K1MGB

Weights + overheadKV cache

Can the RTX 4060 run Spark-X2.6-4B-Base? →

Speed vs quality per quant

Smaller quants run faster and fit in less memory, but lose quality.

Q2_KQ3_K_MQ4_K_MQ5_K_MQ6_KQ8_0F16

Fits in VRAMTight fitCPU offload✕ won't runQuality is typical KL divergence vs FP16, not measured for this model.

Real-world speed of Spark-X2.6-4B-Base

Tokens per second measured by people running it at home. A GPU shows a median once 3 reports agree on the same setup. Reports far from our estimate wait for a review. Measure yours with llama-bench →

No medians yet. Be one of the first 3 to report a setup.

All reports

Run Spark-X2.6-4B-Base locally

--quantQ4_K_Mdefault; pick your GPU to tailor itmradermacher/Spark-X2.6-4B-Base-i1-GGUF

Q4_K_M straight from Hugging Facedocs
ollama run hf.co/mradermacher/Spark-X2.6-4B-Base-i1-GGUF:Q4_K_M

Install Ollama

OpenAI-compatible server with a web UIdocs
llama-server -hf mradermacher/Spark-X2.6-4B-Base-i1-GGUF:Q4_K_M

Downloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.

Install llama.cpp

Download with the lms CLIdocs
lms get mradermacher/Spark-X2.6-4B-Base-i1-GGUF@Q4_K_M

Then load it from the LM Studio app or with `lms load`.

Install LM Studio

Serve an OpenAI-compatible APIdocs
sglang serve XHToken/Spark-X2.6-4B-Base --host 0.0.0.0 --port 30000
Docker Compose (NVIDIA)docs
yaml
services:
  sglang:
    image: lmsysorg/sglang:latest
    command: sglang serve XHToken/Spark-X2.6-4B-Base --host 0.0.0.0 --port 30000
    ports:
      - "30000:30000"
    ipc: host
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]

API on port 30000.

Install SGLang

Chat with the pipeline APIdocs
python
from transformers import pipeline

pipe = pipeline("text-generation", model="XHToken/Spark-X2.6-4B-Base", dtype="auto", device_map="auto")
messages = [{"role": "user", "content": "Explain self-hosting in one paragraph."}]
out = pipe(messages, max_new_tokens=512)
print(out[0]["generated_text"][-1]["content"])

Install Transformers

Quantised versions and downloads

About Spark-X2.6-4B-Base

A little over a month has passed since Spark-X2.5. In the rapidly evolving world of AI agents, that is enough time to rethink, refine, and move the frontier forward. Building on the Spark-X2.5-4B, we introduce Spark-X2.6-4B, the latest generation of our compact, general-purpose language models for on-device deployment. Spark-X2.6-4B retains exactly the same architecture as Spark-X2.5-4B, achieving its capability gains through improved post-training. It delivers substantial improvements in coding, agent, long-context, multilingual, reasoning, etc. These advances enable more capable and practical on-device AI for developers and users tackling real-world tasks. Technical Highlights: - Stronger Coding, Agentic, and Long-Context Capabilities: The models deliver substantial improvements in agentic workflows involving complex tool use, multi-step planning, and software engineering. They achieve state-of-the-art performance among similarly sized open-source models, matching or even surpassing models with approximately 9B parameters. - Controllable Thinking Budget: The models provide three configurable thinking levels—none, low, high—allowing applications to balance reasoning quality, latency, and inference cost dynamically. We recommend using Llama-Factory to fine-tune the model.

Full model card on Hugging Face →

Alternatives to Spark-X2.6-4B-Base

All alternatives →
nvidia logo≈4.2 GB

NVIDIA-Nemotron-3-Nano-4B-BF16 is a small language model (SLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. It responds to user queries and tasks by first generating a reasoning trace and then concluding with a final response.

4B256K ctx nvidia-nemotron-open-model-license
1.9MMar 2026gguf
Qwen logo≈4.0 GB

We introduce the updated version of the Qwen3-4B non-thinking mode, named Qwen3-4B-Instruct-2507, featuring the following key enhancements:

4B256K ctx Apache 2.0
2.7MAug 2025gguf
Qwen logo

Qwen3-4B

Qwen·Chat

≈4.0 GB

Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models.

4B40K ctx Apache 2.0
3.1MApr 2025ollama
Nanbeige logo

Nanbeige4.2-3B

Nanbeige·Chat

≈3.7 GB

Nanbeige4.2-3B is a compact agentic model built on Nanbeige4.2-3B-Base, designed to combine strong agentic behavior with broad reasoning and alignment capabilities. Its Looped Transformer architecture reuses the transformer layers to increase model capacity without adding parameters.

4.2B256K ctx Apache 2.0
89.1KJul 2026gguf
empero-ai logo

Qwen3.8-4B-Distill

empero-ai·Chat

≈4.3 GB

Qwen3.8-4B is a full-parameter distillation of Qwen3.8 2.4T A95B into the Qwen3.5-4B architecture. The student was trained on ~45,000 curated teacher traces from our internal Qwen3.8 distillation datasets — dense chain-of-thought spanning mathematics, general reasoning, and instruction following…

4.7B256K ctx Apache 2.0
112.4KAug 2026gguf
ibm-granite logo

granite-4.1-3b

ibm-granite·Chat

≈3.1 GB

Model Summary: Granite-4.1-3B is a 3B parameter long-context instruct model finetuned from Granite-4.1-3B-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets.

3.4B128K ctx Apache 2.0
436KApr 2026ollama

Spark-X2.6-4B-Base FAQ

How much VRAM does Spark-X2.6-4B-Base need?
About 4.1 GB at Q4_K_M with an 8K context. Q8_0 needs about 5.9 GB and full precision about 9.7 GB. Longer contexts need more memory for the KV cache.
Can Spark-X2.6-4B-Base run on the RTX 4060?
Yes. It fits at Q8_0 (about 5.9 GB), at an estimated 34 tokens/s. It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.
Can I use Spark-X2.6-4B-Base commercially?
Yes. It is released under Apache 2.0. Permissive: use, modify and sell freely; keep the notice and licence text.
What is the context length of Spark-X2.6-4B-Base?
1,048,576 tokens (1M). The full window needs far more memory than an 8K context.
How do I run Spark-X2.6-4B-Base with Ollama or llama.cpp?
With Ollama: ollama run hf.co/mradermacher/Spark-X2.6-4B-Base-i1-GGUF:Q4_K_M. With llama.cpp: llama-server -hf mradermacher/Spark-X2.6-4B-Base-i1-GGUF:Q4_K_M, which starts an OpenAI-compatible server.

Data from Hugging Face, the model card and public leaderboards. Memory and speed figures are estimates; real numbers depend on the runtime, context length and drivers. Always check the licence yourself before commercial use.