Downloads · 30 days
562
17% of all-time downloads
K2-Horizon-MoVA-36B-A4B-MLX-4bit
abenzerps/K2-Horizon-MoVA-36B-A4B-MLX-4bit
K2-Horizon-MoVA-36B-A4B-MLX-4bit is a text generation model from abenzerps. Use it when you need the model to write or continue text. It is set up for mlx. The card lists the license as apache-2.0.
[](https://abenzerps.com/request-a-model/)
All-time downloads
3.3K
Public
Parameters
37.4B
21.1 GB on disk
Likes
4
Public
Repo makeup
Click a slice to open those files.
.safetensors21.1 GB · 100%
Parameter types
How the weights are stored.
U3237.4B · 100%
Try a prompt
Base models
Model card
From the Hugging Face model README
K2-Horizon-MoVA-36B-A4B MLX-4bit
MLX 4-bit conversion of IFM/K2-Horizon-MoVA-36B-A4B, a sparse Mixture-of-Experts model with Mixture-of-Values attention (MoVA), 36B total parameters, and approximately 4B active parameters per token. The source checkpoint supports a native context length of 524,288 tokens (512K).
Benchmarks
Benchmark results reported by IFM for the original K2-Horizon-MoVA-36B-A4B checkpoint.
Release
| Format | Quantization | Size |
|---|---|---|
| MLX safetensors | Affine 4-bit, group size 64 | 21.07 GB |
The MoE and MoVA router weights remain at 8-bit; other eligible weights use affine 4-bit quantization. This is a text-only release; no vision projector is required. The repository includes a source-compatible chat_template.jinja, the custom MLX model implementation required by this architecture, and SHA256SUMS.
Usage
pip install -U mlx-lm
mlx_lm.generate \
--model abenzerps/K2-Horizon-MoVA-36B-A4B-MLX-4bit \
--prompt "Explain why reproducible builds matter." \
--max-tokens 512 --temp 1.0 --top-p 0.95
Use a current mlx-lm release. The source model supports 512K context; usable context length depends on available unified memory and KV-cache settings.
Source
- Model: IFM/K2-Horizon-MoVA-36B-A4B
- Source revision:
05cab0a - License: Apache-2.0
- Checksums: SHA256SUMS
- Task
- Text Generation
- Library
- mlx
- Type
- k2_horizon
- License
- apache-2.0
- Languages
- en
- Created
- Sep 3, 2026
- Updated
- Oct 11, 2026