alt.freeweights.text / Spark-X2.6-4B

From: XHToken <huggingface.co/XHToken>Newsgroups: alt.freeweights.textSubject:

Spark-X2.6-4B

Date: 2026-11-15Commit: 64ca8f2Signed: 2DC28CC7B26A283B ✓

Text generation model with 4.1B parameters. Safetensors weights, 8.24 GB in all. Published by XHToken under Apache-2.0.

4.1B params8.24GBSpark-X2.6-4Bsafetensorspermissive

FactsParams4.1BSize8.24 GBFormatssafetensorsTasktext generationLicenseApache-2.0From sourcepermissive

$ fw get XHToken/Spark-X2.6-4B
# manifest sha256 55cb5a3752f4… · 8.24 GB · 25 files
$ fw verify ./Spark-X2.6-4B XHToken/Spark-X2.6-4B

From sourceHosted (signed)XHToken/Spark-X2.6-4BLast check: 2026-11-15 02:36 UTC

Run at home Get help

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Fine-tune of Spark-X2.6-4B-Base.

Works with

Run24 Fine-tune29 Convert15 Evaluate13 Scan13 Apps58

Runs with

Ranked by fit to this model (safetensors), then by GitHub stars.

vLLM exo Text Generation WebUI

SGLang MLC LLM KTransformers LMCache Xinference LMDeploy mistral.rs Mooncake FlashInfer

Show all 24

AIBrix fastllm ExLlamaV3

Transformers LocalAI tinygrad gpu-unlock Triton Inference Server BentoML CTranslate2 ExLlamaV2 LitServe

Convert with

mergekit TensorRT Model Optimizer

LLM Compressor Neural Compressor AutoRound AQLM GPTQModel ONNX Core ML Tools ExecuTorch

Show all 15

Evaluate with

Opik DeepEval Ragas lm-evaluation-harness OpenCompass SWE-bench

simple-evals TruLens HELM

LightEval

Evaluate Weave

Show all 13

Autoevals

Scan with

promptfoo garak Guardrails NeMo Guardrails Giskard DeepTeam AI Fairness 360 Fairlearn Detoxify MCP Scanner AgentDojo ModelScan

Show all 13

picklescan

Overview Model card Files Use it License Provenance Model tree

Overview

Size: 8.24 GB
Weights: 8.22 GB (safetensors)
Parameters: 4.1B
Formats: safetensors
Task: text generation
License: Apache-2.0 (permissive)
Files: 25, each with a sha256
Publisher: XHToken
Added: 2026-11-15
Section: alt.freeweights.text

Lineage

Original release from the publisher.

Base model: XHToken/Spark-X2.6-4B-Base

Where these bytes come from

Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources

Badge	Host	Last check	Result
From source	huggingface.co	2026-11-15 02:36 UTC	passed

Model card

From the publisher at commit 64ca8f2. Shown as plain text with links; images open at the publisher.

Spark-X2.6-4B: Advancing the Frontier of On-Device Agents for Real-World Tasks#

Slack Discord YouTube dev.to Bluesky X Zhihu WeChat

Note. This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format.

Image: Spark-X2.6 benchmark comparison

Introduction#

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.

Benchmarks#

We evaluate our models and compare them with leading on-device models of similar size across a broad range of tasks, including agent, code, math, general and knowledge. Spark-X2.6-4B achieves best-in-class performance among models of comparable size, while matching or exceeding the performance of larger 9B models on several benchmarks.

Image: all Spark-X2.6 benchmarks

Edge Inference#

Spark-X2.6-4B delivers inference efficiency on par with or better than comparable on-device models, including Qwen3.5-4B and Gemma-E4B. It combines this efficiency with stronger capabilities, making it a practical choice for responsive, resource-conscious on-device AI.

Image: Spark-X2.6 inference

Reasoning Effort Comparison#

We compare high and low reasoning effort across agent, code, math, long-context, and general knowledge tasks. When switching from high to low, Spark-X2.6-1.7B average score decreases by 4.0 points, while its average output length drops by 37.3%. For Spark-X2.6-4B, the average score decreases by 1.7 points, while average output length drops by 22.0%. This flexibility allows developers to balance task performance and generation efficiency according to their application needs.

Image: Spark-X2.6 reasoning_effort

Quickstart#

Transformers#

Install the required version with pip install transformers==4.57.1.

from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "XHToken/Spark-X2.6-4B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    trust_remote_code=True,
    torch_dtype="auto",
    device_map="auto",
)
messages = [{"role": "user", "content": "What is the capital of Anhui Province?"}]
inputs = tokenizer.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    enable_thinking=True,
    return_dict=True,
    return_tensors="pt",
).to(model.device)
outputs = model.generate(
    **inputs,
    max_new_tokens=131072,
    do_sample=True,
    temperature=1.0,
    top_p=0.95,
    top_k=0,
)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

SGLang#

The examples below serve a local Spark-X2.6-4B checkpoint. Set MODEL_PATH to its absolute path before starting a container:

export MODEL_PATH=/absolute/path/to/Spark-X2.6-4B

Install SGLang#

Use the pre-built image that tracks the Spark-X2.6 runtime:

For NVIDIA GPUs:#

docker pull lmsysorg/sglang:nightly-dev-cu13-20260827-20621aa1

For Ascend NPUs:#

# A3 daily build
export SGLANG_IMAGE=quay.io/ascend/sglang:main-cann9.0.0-a3

# A2 daily build (use this instead on A2 hardware)
export SGLANG_IMAGE=quay.io/ascend/sglang:main-cann9.0.0-910b

docker pull "$SGLANG_IMAGE"

Run Inference#

The following commands start an OpenAI-compatible API server configured for a maximum context length of 1,048,576 tokens. This setting requires sufficient device memory; reduce --context-length when necessary.

Server#

NVIDIA GPU:#

docker run --rm -it \
  --gpus '"device=0"' \
  --ipc=host \
  -p 30000:30000 \
  -v "$MODEL_PATH:/root/Spark-X2.6-4B:ro" \
  lmsysorg/sglang:nightly-dev-cu13-20260827-20621aa1 \
  python -m sglang.launch_server \
    --model-path /root/Spark-X2.6-4B \
    --served-model-name spark2.6 \
    --tool-call-parser spark25 \
    --reasoning-parser qwen3 \
    --tp-size 1 \
    --mem-fraction-static 0.8 \
    --context-length 1048576 \
    --chat-template /root/Spark-X2.6-4B/chat_template.jinja \
    --host 0.0.0.0 \
    --port 30000

Ascend NPU:#

docker run -it --rm -e ASCEND_USE_FIA=1 --network=host --ipc=host --shm-size=16g \
    --device=/dev/davinci0 --device=/dev/davinci1 --device=/dev/davinci2 --device=/dev/davinci3 \
    --device=/dev/davinci4 --device=/dev/davinci5 --device=/dev/davinci6 --device=/dev/davinci7 \
    --device=/dev/davinci8 --device=/dev/davinci9 --device=/dev/davinci10 --device=/dev/davinci11 \
    --device=/dev/davinci12 --device=/dev/davinci13 --device=/dev/davinci14 --device=/dev/davinci15 \
    --device=/dev/davinci_manager \
    --device=/dev/devmm_svm \
    --device=/dev/hisi_hdc \
    --volume /usr/local/sbin:/usr/local/sbin \
    --volume /usr/local/Ascend/driver:/usr/local/Ascend/driver \
    --volume /usr/local/Ascend/firmware:/usr/local/Ascend/firmware \
    --volume /etc/ascend_install.info:/etc/ascend_install.info \
    --volume /var/queue_schedule:/var/queue_schedule \
    --volume ~/.cache/:/root/.cache/ \
    --volume "$MODEL_PATH:/root/Spark-X2.6-4B:ro" \
    --entrypoint=python \
    "$SGLANG_IMAGE" \
    -m sglang.launch_server \
      --model-path /root/Spark-X2.6-4B \
      --served-model-name spark2.6 \
      --tool-call-parser spark25 \
      --reasoning-parser qwen3 \
      --tp-size 1 \
      --mem-fraction-static 0.8 \
      --context-length 1048576 \
      --chat-template /root/Spark-X2.6-4B/chat_template.jinja \
      --host 0.0.0.0 \
      --port 30000

Client#

Thinking is enabled by default by both the chat template and the Qwen3 reasoning parser, default reasoning_effort set to "high". You can configure thinking behavior for each request through chat_template_kwargs: set reasoning_effort to "low" or "high", such as "chat_template_kwargs": {"reasoning_effort": "low"}, or set "chat_template_kwargs": {"enable_thinking": false} to disable thinking.

curl -s http://localhost:30000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "spark2.6",
    "messages": [
      {
        "role": "user",
        "content": "安徽的省会在哪里？"
      }
    ],
    "chat_template_kwargs": {"reasoning_effort": "low"},
    "max_tokens": 131072,
    "temperature": 1,
    "top_k": -1,
    "top_p": 0.95,
    "repetition_penalty": 1,
    "presence_penalty": 0,
    "frequency_penalty": 0
  }'

vLLM#

Deploy vLLM#

vLLM provides an official Docker image for NVIDIA GPU deployment:

docker run --rm --gpus all \
  --ipc=host \
  -p 30000:30000 \
  -v "$MODEL_PATH:/models/Spark-X2.6-4B:ro" \
  vllm/vllm-openai:latest \
  --model /models/Spark-X2.6-4B \
  --port 30000 \
  --trust-remote-code \
  --served-model-name spark25 \
  --tensor-parallel-size 1 \
  --gpu-memory-utilization 0.7 \
  --enable-prefix-caching \
  --chat-template /models/Spark-X2.6-4B/chat_template.jinja

For Ascend NPUs, choose an official image for the fastest setup.

Ascend A2:#

export IMAGE=quay.io/ascend/vllm-ascend:nightly-main
docker pull "$IMAGE"

export DEVICE=/dev/davinci0
export MODEL_CACHE="${HOME}/.cache"

mkdir -p "$MODEL_CACHE"

docker run --rm \
    --name vllm-ascend \
    --shm-size=1g \
    --device "$DEVICE" \
    --device /dev/davinci_manager \
    --device /dev/devmm_svm \
    --device /dev/hisi_hdc \
    -v /usr/local/dcmi:/usr/local/dcmi \
    -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
    -v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
    -v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
    -v /etc/ascend_install.info:/etc/ascend_install.info \
    -v "$MODEL_CACHE:/root/.cache" \
    -p 8000:8000 \
    -it "$IMAGE" bash

Ascend A3:#

export IMAGE=quay.io/ascend/vllm-ascend:nightly-main-a3
docker pull "$IMAGE"

export DEVICE0=/dev/davinci0
export DEVICE1=/dev/davinci1
export MODEL_CACHE="${HOME}/.cache"

mkdir -p "$MODEL_CACHE"

docker run --rm \
    --name vllm-ascend \
    --shm-size=1g \
    --device "$DEVICE0" \
    --device "$DEVICE1" \
    --device /dev/davinci_manager \
    --device /dev/devmm_svm \
    --device /dev/hisi_hdc \
    -v /usr/local/dcmi:/usr/local/dcmi \
    -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
    -v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
    -v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
    -v /etc/ascend_install.info:/etc/ascend_install.info \
    -v "$MODEL_CACHE:/root/.cache" \
    -p 8000:8000 \
    -it "$IMAGE" bash

Ascend 950DT:#

export IMAGE=quay.io/ascend/vllm-ascend:nightly-main-a5
docker pull "$IMAGE"

export MODEL_CACHE="${HOME}/.cache"

mkdir -p "$MODEL_CACHE"

docker run --rm \
    --name vllm-ascend \
    --net=host \
    --shm-size=1g \
    --device /dev/davinci0 \
    --device /dev/davinci_manager \
    --device /dev/devmm_svm \
    --device /dev/hisi_hdc \
    -v /usr/local/dcmi:/usr/local/dcmi \
    -v /usr/local/Ascend/driver/tools/hccn_tool:/usr/local/Ascend/driver/tools/hccn_tool \
    -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
    -v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
    -v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
    -v /etc/ascend_install.info:/etc/ascend_install.info \
    -v "$MODEL_CACHE:/root/.cache" \
    -it "$IMAGE" bash

Install the Spark plugin inside the container:

pip install uv
uv venv ~/spark2_6
source ~/spark2_6/bin/activate
git clone https://github.com/XHToken/Spark-plugin.git
cd ./Spark-plugin
uv pip install .

Server#

vllm serve "/models/Spark-X2.6-4B" \
 --port "30000" \
 --trust-remote-code \
 --served-model-name spark25 \
 --tensor-parallel-size 1 \
 --gpu-memory-utilization 0.7 \
 --enable-prefix-caching \
 --chat-template /models/Spark-X2.6-4B/chat_template.jinja

Client#

curl -s http://127.0.0.1:30000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "spark25",
    "messages": [{"role": "user", "content": "What is the capital of Anhui Province?"}],
    "temperature": 1.0, 
    "top_k": -1,
    "top_p": 0.95
  }'

MLX#

Spark-MLX-LLM runs the original Spark-X2.6 Hugging Face checkpoints locally. It supports Apple silicon GPU, Linux CPU, and NVIDIA CUDA on Linux. No GGUF conversion is required.

Installation#

git clone https://github.com/XHToken/Spark-MLX-LLM.git
cd Spark-MLX-LLM

python3 -m venv .venv
source .venv/bin/activate

# Apple silicon
python -m pip install -e .
# Linux CPU
python -m pip install -e '.[cpu]'
# Linux with CUDA 12
python -m pip install -e '.[cuda12]' 
# Linux with CUDA 13
python -m pip install -e '.[cuda13]'

Run Inference#

spark-mlx-generate \
  --device gpu \
  --dtype bfloat16 \
  --model XHToken/Spark-X2.6-4B \
  --prompt "What is the capital of Anhui Province?" \
  --max-tokens 512 \
  --temp 0

llama.cpp#

Note. Native spark2_5 support requires llama.cpp b10828 or later.

Run the model in the CLI:

llama cli -m /path/to/Spark-X2.6-4B-GGUF

Launch the OpenAI-compatible API server:

llama serve -m /path/to/Spark-X2.6-4B-GGUF

Ollama#

Download and install Ollama from ollama.com (v0.34.1 or later).
Run the model:

ollama run SparkLLM/Spark-X2.6-4B

LM Studio#

Download and install LM Studio from lmstudio.ai (0.4.0 or later).
Search for "Spark-X2.6" in LM Studio and download the model.
Load the model and start chatting.

Run with the lms cli#

# List available models
lms ls

# Run chat with Spark-X2.6
lms chat spark-x2.6

Unsloth#

Download Unsloth Studio

Note. The latest Unsloth Studio supports native Spark-X2.6 GGUF inference.

WebGPU (Community)#

Run Spark-X2.6-4B locally in your browser with the community WebGPU demo contributed by Mohd Mahmodi.

Open the WebGPU demo in a WebGPU-compatible browser.
Download the model through the demo and start chatting. Inference runs on your local GPU.

ONNX model files · Source code and validation

Note. This community implementation uses the official Spark-X2.6-4B-INT4 weights and requires no local installation. Tested by the contributor on Windows with an NVIDIA GPU.

Fine-Tuning#

We recommend using Llama-Factory to fine-tune the model.

License#

The Spark-X2.6 model series is licensed under the Apache 2.0 License.

Citation#

If you find our work helpful, feel free to give us a cite.

@misc{sparkx2.6,
    title  = {Spark-X2.6-4B&1.7B: Advancing the Frontier of On-Device Agents for Real-World Tasks},
    author = {SparkLLM Team},
    year   = {2026}
}

Files

25 files · 8.24 GBWeights 8.22 GB (safetensors). Sizes and sha256 from the signed manifest.

Show all 25 files with their sha256

File	Size	sha256
.gitattributes	1.79 KB	b741b8835b5146e8495c608bbbd6de82e8d1d32ca0a1a12f646fee809097b88e
LICENSE	11.3 KB	525c465ab952779d2d360a9ff36f7f49ae018a23d80302ff839a850a77784c36
README.md	15.8 KB	02d5009d9d54004c2c86be0e0d221716c46bdeac4917ace3e1939687a021ec45
chat_template.jinja	5.72 KB	480576bd2ef80726a0dba6406d72ca37f4f4f9ca747c7f1c8d159dbdf40f6929
config.json	2.04 KB	767161ece8ed2891e44345afdf98d29c1cacb964116cb6db7d20dc1c757490ad
configuration_spark.py	4.33 KB	02218597240490f490659b184052db940b08163ff9c3af7b7e59323f6f963722
generation_config.json	281 B	804a5276a9d237600688267dd2bf6f9d4129ec21faf417f870d5d52b32ad9e6c
images/.DS_Store	6.15 KB	d65165279105ca6773180500688df4bdc69a2c7b771752f0a46ef120b7fd8ec3
images/all_benchmark.svg	45.9 KB	0f766e6d6acc297031b56b104cb29c78ce235d4669c86dd76392792a41871983
images/benchmark.svg	51.4 KB	7efb3e04fc5c8e356727ba0ca1c401c16c3a3b007e3e942735e61c414bf0b3da
images/inference.svg	25.7 KB	a4e2a93633aebdc81037d42c649126ebefb2b52051fc639681a60e1f363456b4
images/reasoning_effort.svg	13.6 KB	8d0b3f424a0cedc1703b9911ee715e19a1aaad468a686d68ad4ebfa13a97d5fa
images/xhtoken-wechat.jpg	160 KB	0c3fbd25c987237b383461fdaa96bf6881af615c23ba3baf19e51a7879354ddb
merges.txt	1.57 MB	b06fc88940ae6a20f71142f06304c12547f58ab52c175e56af12a38c3dc62f36
model-00001-of-00005.safetensors	1.98 GB	9ba8a59a47ce4197d5c749c76c04272e0b4b880747b31b9167a82d8757e90665
model-00002-of-00005.safetensors	1.99 GB	f0cad973670dda858a177d62a1845f9ec3c3e2a7a1cdbd34f98ab6789fbd452d
model-00003-of-00005.safetensors	1.99 GB	479c8aecd24d1d4c6c2243541b36acc6a84df4c0706a7ccaacda748eac46b8e8
model-00004-of-00005.safetensors	1.99 GB	766cd843c58676b0be533b465368a76f574f36bea074f4d16f7e37d910423d08
model-00005-of-00005.safetensors	262 MB	e2cb22cc6aa1dccdc30042170841058a1a8cffb7e9eb57356c86ae14fc4ca0e3
model.safetensors.index.json	24.1 KB	77a11d49b279b5e44fcb1f2453f71a8023237f93992fe5275d842e149d896edd
modeling_spark.py	19.4 KB	9cf0d1ad2b54b9f7088792779ddd8b5cb4d4fe63bfea054da7f2dd16362bcaf4
special_tokens_map.json	156 B	48a257e3f5f20a004e7c30482a70ae19ae7d9cf713144f2a1394ba20af8c3cb5
tokenizer.json	10.1 MB	710cce15cf3565674c499f9413997c6e8101f2bdd96245cff8f0311fb501248c
tokenizer_config.json	5.84 KB	57b385dcc2ad031aef4a3786e97cd9a832ecb839cb27ed2cd829d627c90e1bfd
vocab.json	3.03 MB	9f42cb7e7af5afe27cae2337d272d60b5659e10143ef4bb952f642b847916fe1

VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.

Upstream commit	Date	Contents
64ca8f2a29 pinned now	2026-11-15 02:36 UTC	25 files · 8.24 GB

Use it

$ fw get XHToken/Spark-X2.6-4B
# manifest sha256 55cb5a3752f4… · 8.24 GB · 25 files
$ fw verify ./Spark-X2.6-4B XHToken/Spark-X2.6-4B

fw.pyChecks the signature and every sha256, tries each source in order, resumes.

$ python3 fw.py get https://freeweights.allternit.com/manifests/XHToken/Spark-X2.6-4B.json --dir ./Spark-X2.6-4B
$ python3 fw.py verify ./Spark-X2.6-4B ./XHToken__Spark-X2.6-4B.json

hfThe Hugging Face CLI with one setting. Same repo id.

$ HF_ENDPOINT=https://freeweights.allternit.com hf download XHToken/Spark-X2.6-4B --revision 64ca8f2a295b8ceb2757b9d4749a6558c4759e5b --local-dir ./Spark-X2.6-4B

transformersLoad the weights in Python.

import os
os.environ["HF_ENDPOINT"] = "https://freeweights.allternit.com"
from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("XHToken/Spark-X2.6-4B", revision="64ca8f2a295b8ceb2757b9d4749a6558c4759e5b")
model = AutoModelForCausalLM.from_pretrained("XHToken/Spark-X2.6-4B", revision="64ca8f2a295b8ceb2757b9d4749a6558c4759e5b", device_map="auto")

vLLMServe an OpenAI-compatible API.

$ HF_ENDPOINT=https://freeweights.allternit.com vllm serve XHToken/Spark-X2.6-4B --revision 64ca8f2a295b8ceb2757b9d4749a6558c4759e5b

Run at home

License

Name: Apache License 2.0
Id: Apache-2.0
Class: Permissive
Redistribution: yes
Listing: Hosted (signed)
Decided: License as declared by the publisher. No person read it. 2026-11-15
Text in: LICENSE
Text sha256: 525c465ab952779d2d360a9ff36f7f49ae018a23d80302ff839a850a77784c36
Terms gate: No extra terms beyond the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From: XHToken <huggingface.co/XHToken>Newsgroups: alt.freeweights.textSubject: Spark-X2.6-4BDate: 2026-11-15Commit: 64ca8f2Signed: 2DC28CC7B26A283B ✓

Upstream: XHToken/Spark-X2.6-4B
Pinned commit: 64ca8f2a295b8ceb2757b9d4749a6558c4759e5b
Manifest: Spark-X2.6-4B.json · .minisig
Manifest sha256: 55cb5a3752f41ef02886ead45776c62e534b1cddb5d66d46c191357dc7ebe186
Signature: minisign, Ed25519, key 2DC28CC7B26A283B
Created: 2026-11-15 02:36 UTC
Generator: freeweights-manifest/0.1.0
Takedowns: None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/XHToken/Spark-X2.6-4B.json
$ curl -O https://freeweights.allternit.com/manifests/XHToken/Spark-X2.6-4B.json.minisig
$ minisign -Vm Spark-X2.6-4B.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq
$ shasum -a 256 <file>   # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

Parent

XHToken/Spark-X2.6-4B-Base · this model is a finetune of it

4 models in the catalog name this one as their parent.

Fine-tunes: 1
Quantizations: 3

Fine-tunes 1

JuneoAI/Laplace-1-Vector8.24 GB

Quantizations 3

bartowski/Spark-X2.6-4B-GGUF62.6 GB
mlx-community/Spark-X2.6-4B-OptiQ-4bit3.05 GB
FrostWolfAI/Spark-X2.6-4B-abliterated8.24 GB

Get it

Download

$ HF_ENDPOINT=https://freeweights.allternit.com hf download XHToken/Spark-X2.6-4B

Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest

minisign, key 2DC28CC7B26A283B. Spark-X2.6-4B.json. Provenance. Takedown history: none.

Source

huggingface.co/XHToken/Spark-X2.6-4B

Variants

3 quantized or adapter builds of Spark-X2.6-4B. Each has its own license check and page.

SubjectSizeParamsLicenseDate

Spark-X2.6-4B-GGUFHosted (signed) From: bartowski in alt.freeweights.text62.6 GB—Permissive2026-11-15

Spark-X2.6-4B-OptiQ-4bitHosted (signed) From: mlx-community in alt.freeweights.text3.05 GB4.1BPermissive2026-11-15 Spark-X2.6-4B-abliteratedHosted (signed) From: FrostWolfAI in alt.freeweights.abliterated8.24 GB4.1BPermissive2026-11-15

Related models

All of alt.freeweights.text →

From: laion <huggingface.co/laion>Subject: a2-rl-e2egit_large-50-32B

From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_e2egit-large-15-8B

From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_e2egit-v2-10-8B

From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_pymethods2test-large-80-8B

From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_unitsyn-python-large-10-8B

From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_mix_h2_language_proportional-65-8B

Spark-X2.6-4B

Works with

Runs with

Fine-tune with

Convert with

Evaluate with

Scan with

Works with

Overview

Model card

Spark-X2.6-4B: Advancing the Frontier of On-Device Agents for Real-World Tasks#

Introduction#

Benchmarks#

Edge Inference#

Reasoning Effort Comparison#

Quickstart#

Transformers#

SGLang#

Install SGLang#

For NVIDIA GPUs:#

For Ascend NPUs:#

Run Inference#

Server#

NVIDIA GPU:#

Ascend NPU:#

Client#

vLLM#

Deploy vLLM#

Ascend A2:#

Ascend A3:#

Ascend 950DT:#

Server#

Client#

MLX#

Installation#

Run Inference#

llama.cpp#

Ollama#

LM Studio#

Run with the lms cli#

Unsloth#

WebGPU (Community)#

Fine-Tuning#

License#

Citation#

Files

Use it

License

Provenance

How to verify

Model tree

Variants

Related models