IFM/K2-Horizon-MoVA-36B-A4B-GGUF · Hugging Face
Heat trend
Collecting trend data
The percentage is based on available heat signal, not comment count or independent people.
Hugging Face model activity is surfacing — worth tracking for capability changes, ecosystem impact, and availability.
The IFM/K2-Horizon-MoVA-36B-A4B-GGUF model, available on Hugging Face, demonstrates "Frontier-class results at 4B active parameters." It surpasses open-weight dense models (around 30B model size) and MoE models up to 15 times its size on agentic and reasoning benchmarks. Furthermore, it competes effectively against closed frontier models, as detailed in its Benchmark Results. More sizes are expected to be uploaded for this model.
more sizes (probably still uploading):
https://huggingface.co/IFM/K2-Horizon-32B-GGUF
https://huggingface.co/IFM/K2-Horizon-7B-GGUF
https://huggingface.co/IFM/K2-Horizon-3.7B-GGUF
https://huggingface.co/IFM/K2-Horizon-0.9B-GGUF
from IFM:
K2-Horizon-MoVA-36B-A4B is the sparse member of the K2-Horizon family: a Mixture-of-Experts model with Mixture-of-Values attention (MoVA) that stores 36B parameters and runs 4B per token. We have released the final checkpoint; intermediate checkpoints, along with the data and the training code, will be released.
K2-Horizon-MoVA-36B-A4B Highlights
- Frontier-class results at 4B active parameters. On agentic and reasoning benchmarks it outscores open weight dense (approximately 30B model size) and MoE models up to 15× its size; and also performs competitively against closed frontier models (see Benchmark Results ).
- 512K context. Native 524,288-token context from the midtraining stages onward.
- Intermediate checkpoints. Intermediate checkpoints will be released so capability changes can be studied across training rather than at a single checkpoint.
- Fully open. Training data/recipe and the training code will be made public.
collection: https://huggingface.co/collections/IFM/k2-horizon