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.

inclusionAI logo

ZwZ2-4B

inclusionAI·Instruct·released 15 Oct 2026

Unknown licence4.5B256K ctx ≈4.2 GB
Your GPU
ollama run hf.co/inclusionAI/ZwZ2-4B-GGUF:Q4_K_M
All 6 runners ↓

ZwZ2-4B is the Qwen3.5-4B-based model in Zooming-without-Zooming-v2 (ZwZ2). It learns to connect visual evidence across regions through comparison, aggregation, matching, and multi-step reasoning. Training uses full images; inference requires no zooming tools.

Notify me when

How much VRAM does ZwZ2-4B need?

At Q4_K_M ZwZ2-4B needs about 4.2 GB with an 8K context (4.2 GB at Q4, 10 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)

Quantisation8K16K32K64K128K256K
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 loss1.6 GB file1.0 GB≈3.2 GB
Q3_K_MNoticeable lossSmall, some quality lossNoticeable loss1.9 GB file1.0 GB≈3.5 GB
Q4_K_MrecommendedSmall lossRecommended default for local useSmall loss2.5 GB file1.0 GB≈4.2 GB
Q5_K_MNear-losslessNear-lossless for most tasksNear-lossless3.1 GB file1.0 GB≈4.7 GB
Q6_KNear-losslessPractically losslessNear-lossless3.4 GB file1.0 GB≈5.1 GB
Q8_0LosslessLossless in practice, twice the size of Q4Lossless4.2 GB file1.0 GB≈5.9 GB
FP16 / BF16LosslessFull precision, as releasedLossless7.8 GB file1.0 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(2.5 GB weights + 1.0 GB KV cache + 0.6 GB overhead). Pick a GPU above to see how much context fits.

0102030401K2K4K8K16K32K64K128K256KGB

Weights + overheadKV cache

Can the RTX 4060 run ZwZ2-4B? →

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 ZwZ2-4B

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 ZwZ2-4B locally

--quantQ4_K_Mdefault; pick your GPU to tailor itinclusionAI/ZwZ2-4B-GGUF

Q4_K_M straight from Hugging Facedocs
ollama run hf.co/inclusionAI/ZwZ2-4B-GGUF:Q4_K_M

Install Ollama

OpenAI-compatible server with a web UIdocs
llama-server -hf inclusionAI/ZwZ2-4B-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 inclusionAI/ZwZ2-4B-GGUF@Q4_K_M

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

Install LM Studio

Serve an OpenAI-compatible APIdocs
vllm serve inclusionAI/ZwZ2-4B
Docker Compose (NVIDIA)docs
yaml
services:
  vllm:
    image: vllm/vllm-openai:latest
    command: inclusionAI/ZwZ2-4B
    ports:
      - "8000:8000"
    ipc: host
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]

API on port 8000.

Install vLLM

Serve an OpenAI-compatible APIdocs
sglang serve inclusionAI/ZwZ2-4B --host 0.0.0.0 --port 30000
Docker Compose (NVIDIA)docs
yaml
services:
  sglang:
    image: lmsysorg/sglang:latest
    command: sglang serve inclusionAI/ZwZ2-4B --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

Ask about an imagedocs
python
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="inclusionAI/ZwZ2-4B", dtype="auto", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
            {"type": "text", "text": "Describe this image."},
        ],
    }
]
out = pipe(text=messages, max_new_tokens=256)
print(out[0]["generated_text"][-1]["content"])

Install Transformers

Quantised versions and downloads

RepositoryFormatQuantSizeDownloads
bartowski/ZwZ2-4B-GGUFGGUFIQ2_M1.6 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFQ2_K1.7 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFIQ3_XXS1.9 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFQ3_K_S2.0 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFIQ3_XS2.0 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFQ3_K_M2.1 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFQ3_K_L2.2 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFIQ3_M2.3 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFIQ4_XS2.4 GB3.3K
bartowski/ZwZ2-4B-GGUFGGUFQ4_K_S2.5 GB3.3K

About ZwZ2-4B

ZwZ2-4B is the Qwen3.5-4B-based model in Zooming-without-Zooming-v2 (ZwZ2). It learns to connect visual evidence across regions through comparison, aggregation, matching, and multi-step reasoning. Training uses full images; inference requires no zooming tools. On 25 benchmarks across six domains, it improves over its base model on every benchmark, increasing overall accuracy from 49.41 to 54.79 (+5.38 points). Gains extend from fine-grained perception to spatial reasoning, long-context understanding, multimodal agents, and general visual reasoning. EvidenceConnect builds verified multi-region questions from SA-1B images: • Group 3–7 related regions and generate questions using privileged zoomed views. • Check answers through cross-model agreement and drop-one region tests. • Convert verified questions to four-option tasks and train with GRPO, without an additional supervised fine-tuning stage. All three model sizes use approximately 20K training examples. The policy sees full images with colored region outlines; privileged crop layouts are used only for data construction. See ZwZ2-RL-VQA for the data schema and synthesis process. Editable vector PDF. The figure shows eight representative benchmarks; the complete results follow.

Full model card on Hugging Face →

Other ZwZ2 models

Whole family →
inclusionAI logo
ZwZ2-9BinclusionAI
9.4BCheck licence≈7.1 GB↓ 1.5K
inclusionAI logo
ZwZ2-27BinclusionAI
27.4BCheck licence≈19 GB↓ 1.5K

Alternatives to ZwZ2-4B

All alternatives →
Qwen logo

Qwen3.5-4B

Qwen·Vision

≈4.3 GB

Over recent months, we have intensified our focus on developing foundation models that deliver exceptional utility and performance.

4.7B256K ctx Apache 2.0
4.3MFeb 2026ollama
Qwen logo≈4.3 GB

Meet Qwen3-VL — the most powerful vision-language model in the Qwen series to date.

4.4B256K ctx Apache 2.0
2.8MOct 2025ollama
Qwen logo

Qwen3.5-4B-Base

Qwen·Vision

≈4.3 GB

Over recent months, we have intensified our focus on developing foundation models that deliver exceptional utility and performance.

4.7B256K ctx Apache 2.0
496.7KFeb 2026gguf
h2oai logo

h2o-lightning-4b

h2oai·Vision

≈4.2 GB

H2O-Lightning-4B is a 4B-parameter decision model from H2O.ai, built on Qwen/Qwen3.5-4B. It answers typed decision questions about a record (a document, a ticket, a policy, a conversation) and about images that come with the record (photos, screenshots, scanned documents, charts).

4.5B256K ctx Apache 2.0
26.1KOct 2026gguf
tencent logo

WeVisOne-VL-4B

tencent·Vision

≈4.5 GB

WeVisOne-VL (WeChat Vision All-in-One) is a unified vision-language model developed by WeChat Vision, Tencent Inc. Built upon Qwen3-VL, it unifies object-level perception (open-vocabulary object detection, referring localization, object counting, object pointing, and OCR) and pixel-level perception…

4.8B256K ctx Apache 2.0
36.4KOct 2026gguf
google logo

medgemma-1.5-4b-it

google·Vision

≈4.1 GB

* License: The use of MedGemma is governed by the Health AI Developer Foundations terms of use. MedGemma has not been evaluated or optimized for multi-turn applications.

4.3B128K ctx health-ai-developer-foundations
183.5KJan 2026ollama

ZwZ2-4B FAQ

How much VRAM does ZwZ2-4B need?
About 4.2 GB at Q4_K_M with an 8K context. Q8_0 needs about 6.2 GB and full precision about 10 GB. Longer contexts need more memory for the KV cache.
Can ZwZ2-4B run on the RTX 4060?
Yes. It fits at Q8_0 (about 5.9 GB), at an estimated 31 tokens/s. It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.
Can I use ZwZ2-4B commercially?
Unclear. It is released under Unknown licence. No licence stated: assume all rights reserved. Check the licence on the model card before using it in a product.
What is the context length of ZwZ2-4B?
262,144 tokens (256K). The full window needs far more memory than an 8K context.
How do I run ZwZ2-4B with Ollama or llama.cpp?
With Ollama: ollama run hf.co/inclusionAI/ZwZ2-4B-GGUF:Q4_K_M. With llama.cpp: llama-server -hf inclusionAI/ZwZ2-4B-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.