Qwythos-9B-Claude-Mythos-5-1M MLX (8bit, unlocked)
A conversion of empero-ai/Qwythos-9B-Claude-Mythos-5-1M to Apple's MLX format, quantized to 8bit.
- Conversion tool: mlx-maker /
mlx-vlm - Source model:
empero-ai/Qwythos-9B-Claude-Mythos-5-1M - Base model:
Qwen/Qwen3.5-9B - License: Apache-2.0 (inherited from the source model)
About the chat template
Modified variant: the forced identity baked into chat_template.jinja (the Qwythos self-identification) has been removed, and the user's system prompt is now respected. When no system prompt is provided, a neutral default "You are a helpful AI assistant." is used. The weights are unchanged and identical to the same-quantization base variant.
Usage (mlx-vlm)
uv add mlx-vlm
from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
model, processor = load("ToPo-ToPo/Qwythos-9B-Claude-Mythos-5-1M-mlx-8bit-unlocked")
messages = [{"role": "user", "content": "Hello"}]
prompt = apply_chat_template(processor, model.config, messages)
print(generate(model, processor, prompt, max_tokens=256))
This repository is a derivative work distributed under Apache-2.0. The original attribution and license are inherited as noted above.