README.md

# K2-Horizon-MoVA-36B-A4B MLX

MLX conversions of [IFM/K2-Horizon-MoVA-36B-A4B](https://huggingface.co/IFM/K2-Horizon-MoVA-36B-A4B), a sparse Mixture-of-Experts model with Mixture-of-Values attention (36B total / 4B active parameters).

## Available Formats

| Format | Size | Quality | Use Case |
|--------|------|---------|----------|
| **oQ4e** | ~21 GB | ~uniform 6-bit quality | Best quality-per-GB |
| **6-bit** | ~28 GB | High | Quality-focused, fits 40+ GB |
| **8-bit** | ~40 GB | Near-lossless | Reference quality, 64 GB+ |

## Quickstart

```bash
pip install -U mlx-lm

# Generate
python3 -m mlx_lm.generate \
  --model hermitdave/K2-Horizon-MoVA-36B-A4B-MLX-8bit \
  --prompt "Explain why long-context evaluation is difficult." \
  --max-tokens 512 --temp 1.0 --top-p 0.95
```

## Reasoning

K2-Horizon is a reasoning model. Always use `reasoning_effort="high"` for best results:

```python
from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")
response = client.chat.completions.create(
    model="hermitdave/K2-Horizon-MoVA-36B-A4B-MLX-8bit",
    messages=[{"role": "user", "content": "Explain quantum entanglement."}],
    extra_body={"chat_template_kwargs": {"reasoning_effort": "high"}},
)
print("Reasoning:", getattr(response.choices[0].message, "reasoning_content", None))
print("Answer:", response.choices[0].message.content)
```

## Benchmark Results

| Benchmark | K2-Horizon-MoVA-36B-A4B |
|-----------|------------------------|
| tau3-Banking (Agentic tool use) | **26.8** |
| Terminal-Bench 2.1 (Agentic terminal use) | **58.6** |
| GPQA Diamond (Graduate-level science QA) | 80.8 |
| AA-LCR (Long-context reasoning) | 66.3 |

Scores in %. See [model card](https://huggingface.co/IFM/K2-Horizon-MoVA-36B-A4B) for full results.

## Citation

```bibtex
@misc{k2horizon2026,
  title  = {Introducing K2 Horizon: Frontier Performance, Radically Open},
  author = {{IFM Team}},
  year   = {2026},
  url    = {https://ifm.ai/blog/k2/},
}
```

## License

Apache-2.0 (same as upstream).