Skip to content
Other

laya-cua

a compact router for computer-use agents

2 min readupdated

laya-cua is a 421M model built for UI control and integration into Computer-Use Agent pipelines. Based on ModernBERT-large, it takes a text description of the screen, recent step history, and a task, then predicts the next action type and estimates the probability that the task is complete. Rather than generating text inputs or click coordinates, the model serves as a fast decision-making classifier. It is a solid fit for developers building autonomous agents and researchers working on UI automation.

Key features

  • Classification accuracy: reaches 80.8% accuracy in predicting action types on the test dataset.
  • Compact footprint: weighs in at just 421M parameters and runs on standard hardware.
  • Text-only input: processes text descriptions of the screen instead of heavy screenshots.
  • Single-pass inference: determines both the action type and task status in one forward pass.

What you can do

  • Integrate into agent pipelines: run it as a fast router alongside an LLM to pick the next action.
  • Track task completion: estimate the probability that the current on-screen task is finished.

Example

import laya
from huggingface_hub import snapshot_download

ACTIONS = {
  "click": "click or tap a UI element", "type": "type text into a focused field",
  "key": "press a key or shortcut", "scroll": "scroll the view",
  "navigate": "open a URL or switch app or tab", "move": "move the mouse without clicking (hover)",
  "wait": "wait for the page or app to load", "drag": "drag and drop or select by dragging",
  "done": "the task is already complete, stop",
}
QUESTIONS = {
  "action_type": {"type": "choice", "instructions": "Which action type should the agent take next?", "criteria": ACTIONS},
  "task_done":   {"type": "noul",   "instructions": "Is the task already fully complete on screen?"},
}
agent = laya.Agent(snapshot_download("sodan/laya-cua"), device="cuda")
state = {"task": "Open the settings app and turn on dark mode",
         "screen": "Settings window. Sidebar: Network, Display, Sound. Display page shows Theme: Light.",
         "history": ["opened Settings"]}
res = agent.predict(state, QUESTIONS)
print(res["answers"]["action_type"]["choice"], res["answers"]["task_done"]["noul"])
Read moreDetails
Benchmarks: action_type accuracy 0.8084, task_done accuracy 0.9611.
Limitations: trained on AgentNet text descriptions (primarily Ubuntu desktop tasks); performance on mobile interfaces has not been tested.
Base model: convaiinnovations/laya (ModernBERT-large encoder).
Discuss Telegram 0 39

Will it run?

Runs in F16

Fits without quantisation.

  • F16weights 0.8 GB · ready files1.4 GB
  • FP8 / INT8recommendedweights 0.4 GB0.9 GB
  • 4-bitweights 0.2 GB0.7 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · F16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # sodan/laya-cua
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download sodan/laya-cua --local-dir ~/models/laya-cua
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to laya-cua
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of laya (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 1All on Hugging Face

About the model

sodan/laya-cua
Author
sodan
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

390

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments