abenzerps/Holotron4-30B-A3B-MLX

Holotron4-30B-A3B MLX

MLX quantizations of Hcompany/Holotron4-30B-A3B, a 30B hybrid NemotronH MoE vision-language model (VLM) for Computer Use, tool-driven work, and agentic workflows.

Upstream benchmarks

Holotron4 improves over its base model, Nemotron 3 Nano Omni, on GUI workflows and in environments with MCP tools, APIs, or code sandboxes. Gains are absolute percentage points.

Benchmark	Interface	Nemotron 3 Nano Omni	Holotron4-30B-A3B	Gain
OSWorld	GUI	21.0	76.3	+55.3
OSWorld 2.0	GUI and code	0.2	7.9	+7.7
AutomationBench	MCP	19.4	35.6	+16.2
PinchBench	Terminal	84.7	88.6	+3.9
ALE (Linux, code)	Terminal	0.6	8.5	+7.9

MLX Files

Quantization	File	Size
4-bit	Holotron4-30B-A3B-MLX-4bit	18.30 GB
6-bit	Holotron4-30B-A3B-MLX-6bit	25.82 GB
8-bit	Holotron4-30B-A3B-MLX-8bit	33.33 GB

Usage with MLX-VLM

Installation

pip install -U mlx-vlm

Python API

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load 4-bit (or subfolder="Holotron4-30B-A3B-MLX-6bit", subfolder="Holotron4-30B-A3B-MLX-8bit")
model_path = "abenzerps/Holotron4-30B-A3B-MLX"
subfolder = "Holotron4-30B-A3B-MLX-4bit"

model, processor = load(model_path, subfolder=subfolder)
config = load_config(model_path, subfolder=subfolder)

prompt = "Describe the user interface elements shown in this screenshot."
image = ["screenshot.png"]

formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=len(image)
)

output = generate(model, processor, formatted_prompt, image, verbose=True)
print(output)

Command Line Interface

# 4-bit
python -m mlx_vlm.generate \
    --model abenzerps/Holotron4-30B-A3B-MLX --subfolder Holotron4-30B-A3B-MLX-4bit \
    --image screenshot.png \
    --prompt "What action should be taken next to achieve the user goal?"

# 6-bit
python -m mlx_vlm.generate \
    --model abenzerps/Holotron4-30B-A3B-MLX --subfolder Holotron4-30B-A3B-MLX-6bit \
    --image screenshot.png \
    --prompt "What action should be taken next to achieve the user goal?"

# 8-bit
python -m mlx_vlm.generate \
    --model abenzerps/Holotron4-30B-A3B-MLX --subfolder Holotron4-30B-A3B-MLX-8bit \
    --image screenshot.png \
    --prompt "What action should be taken next to achieve the user goal?"

Source

• Base model: Hcompany/Holotron4-30B-A3B
• License: NVIDIA Open Model Agreement
• Checksums: SHA256SUMS