---
base_model: deepreinforce-ai/Ornith-1.0-35B
license: mit
license_link: https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B/blob/main/LICENSE
library_name: mlx
pipeline_tag: image-text-to-text
base_model_relation: quantized
tags:
- mlx
- mlx-vlm
- vlm
- moe
- quantized
---

# Ornith-1.0-35B-6bit

6-bit (group size 64, **6.622 bits/weight**) MLX quantization of
[deepreinforce-ai/Ornith-1.0-35B](https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B),
produced with `mlx-vlm` 0.6.3. Full multimodal: the vision encoder is preserved and
quantized alongside the language model. For Apple Silicon. Runs in `mlx-vlm` or any MLX app.

## Conversion note (MoE expert fusion)

Ornith stores its 256 MoE experts unfused (per-expert), but mlx-vlm's `qwen3_5_moe` loader expects
them fused/batched. A `sanitize` monkeypatch was required to stack the experts before conversion; without it the
conversion failed. This is a standard mlx-vlm 6-bit quant.

## Usage

```bash
uvx --from mlx-vlm mlx_vlm.generate \
  --model mlx-community/Ornith-1.0-35B-6bit --image image.png \
  --prompt "Describe this image." --max-tokens 512
```

```python
from mlx_vlm import load, generate
model, processor = load("mlx-community/Ornith-1.0-35B-6bit")
```

## Conversion check

Smoke-tested after conversion (`mlx_vlm.generate` on an image): coherent — correctly read an
evaluation bar chart, no repetition loop. **102.4 tok/s** generation, **964.8 tok/s** prompt,
**peak 31.1 GB** on a Macbook Pro M5 Max 128GB 40 GPU.

Refer to the original model card for architecture, benchmarks, license, and intended use.