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.

KORMo-VL logo

KORMo-Web-11B

KORMo-VL·Instruct·released 14 Nov 2026

Apache 2.011.2B128K ctx ≈8.4 GB
Your GPU
vllm serve KORMo-VL/KORMo-Web-11B
All 2 runners ↓

KORMo-Web-11B is a general-purpose bilingual AI model built on the KORMo foundation model family, developed by KAIST MLP Lab starting from a language model trained entirely from scratch.

Notify me when

How much VRAM does KORMo-Web-11B need?

At Q4_K_M KORMo-Web-11B needs about 8.4 GB with an 8K context (8.4 GB at Q4, 24 GB at full precision). Pick your GPU to see what fits and how fast it runs.

--vram
any
--gpus
--quant

Needs: ≈8.4 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)

Quantisation4K8K16K32K64K128K
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 loss3.4 GB 1.3 GB≈5.3 GB
Q3_K_MNoticeable lossSmall, some quality lossNoticeable loss5.1 GB 1.3 GB≈7.1 GB
Q4_K_MrecommendedSmall lossRecommended default for local useSmall loss6.3 GB 1.3 GB≈8.4 GB
Q5_K_MNear-losslessNear-lossless for most tasksNear-lossless7.4 GB 1.3 GB≈9.5 GB
Q6_KNear-losslessPractically losslessNear-lossless8.6 GB 1.3 GB≈11 GB
Q8_0LosslessLossless in practice, twice the size of Q4Lossless11 GB 1.3 GB≈13 GB
FP16 / BF16LosslessFull precision, as releasedLossless21 GB 1.3 GB≈24 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.

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 ≈8.4 GB(6.3 GB weights + 1.3 GB KV cache + 0.8 GB overhead). Pick a GPU above to see how much context fits.

01020301K2K4K8K16K32K64K128KGB

Weights + overheadKV cache

Can the Apple M4 (16 GB) run KORMo-Web-11B? →

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 KORMo-Web-11B

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 KORMo-Web-11B locally

These commands load the original weights from Hugging Face.

Serve an OpenAI-compatible APIdocs
vllm serve KORMo-VL/KORMo-Web-11B
Docker Compose (NVIDIA)docs
yaml
services:
  vllm:
    image: vllm/vllm-openai:latest
    command: KORMo-VL/KORMo-Web-11B
    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

Ask about an imagedocs
python
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="KORMo-VL/KORMo-Web-11B", 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

About KORMo-Web-11B

KORMo-Web-11B is a general-purpose bilingual AI model built on the KORMo foundation model family, developed by KAIST MLP Lab starting from a language model trained entirely from scratch. Unlike our Qwen-based KORMo-Web-4B and 9B models, which focus on maximizing web task performance, KORMo-Web-11B is designed to preserve its original LLM and VLM capabilities while learning to act as a GUI agent. It can understand Korean and English text, interpret images, and complete real browser tasks by visually navigating webpages—all within a single model. • Built on Our Own Foundation Model: Developed from the KORMo-10B language model, trained from scratch by KAIST MLP Lab, and extended through KORMo-VL into a web agent. • Three Capabilities in One Model: Designed to retain general LLM reasoning and VLM understanding while adding GUI interaction capabilities, rather than optimizing exclusively for web navigation. • 10M+ Atomic-Level Web Pretraining Samples: Large-scale Korean and English web understanding training, covering individual UI elements, their locations, layouts, functions, and states. • 100K+ Web Interaction Trajectories: Bilingual browser demonstrations for learning real-world, multi-step web tasks. • Open and Commercially Usable: Model weights and supporting resources are openly released under Apache-2.0, allowing commercial use, modification, and redistribution subject to the license terms.

Full model card on Hugging Face →

Alternatives to KORMo-Web-11B

All alternatives →
Qwen logo

Qwen3.5-9B

Qwen·Vision

≈7.2 GB

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

9.7B256K ctx Apache 2.0
6.4MFeb 2026ollama
Qwen logo≈6.8 GB

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

8.8B256K ctx Apache 2.0
6.2MOct 2025ollama
Zyphra logo

ZAYA1-VL-8B

Zyphra·Vision

≈6.6 GB

ZAYA1-VL-8B is a vision-language model (VLM) built upon Zyphra's ZAYA1-8B LLM. It has state-of-the-art performance among VLMs for its size and inference efficiency.

9.7B32K ctx Apache 2.0
337.9KApr 2026gguf
omlab logo

VLX-Seek-1.5-10B

omlab·Vision

≈7.4 GB

VLX-Seek-1.5-10B is the open-source 10B model in the VLX-Seek 1.5 family, designed for fine-grained perception and visual grounding in embodied scenarios.

10B256K ctx Apache 2.0
88.2KJul 2026gguf
Qwen logo

Qwen3.5-9B-Base

Qwen·Vision

≈7.2 GB

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

9.7B256K ctx Apache 2.0
264.8KFeb 2026gguf
black-forest-labs logo

flux-3-image-klein-pup

black-forest-labs·Vision

≈7.1 GB

FLUX 3 Image Klein Prompt Upsampler is a 9B vision-language model that expands a short image request into the detailed prompt FLUX 3 Image Klein works best with. It takes the request, up to 10 reference images and an optional list of bounding boxes, and returns one prompt.

9.4B256K ctx Apache 2.0
39.6KOct 2026gguf

KORMo-Web-11B FAQ

How much VRAM does KORMo-Web-11B need?
About 8.4 GB at Q4_K_M with an 8K context. Q8_0 needs about 13 GB and full precision about 24 GB. Longer contexts need more memory for the KV cache.
Can KORMo-Web-11B run on the RTX 3060 12GB?
Yes. It fits at Q6_K (about 11 GB), at an estimated 21 tokens/s. It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, Arc B580.
Can I use KORMo-Web-11B 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 KORMo-Web-11B?
131,072 tokens (128K). The full window needs far more memory than an 8K context.
How do I run KORMo-Web-11B locally?
With vLLM: vllm serve KORMo-VL/KORMo-Web-11B.

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.