The 5-bit quantized version of the IBM Granite-4.0-H-Tiny model, optimized for Apple silicon chips. It uses a hybrid architecture of Mamba-2 and soft attention, combined with a Mixture of Experts model (MoE), to achieve efficient inference while maintaining high quality.

Natural Language ProcessingMlxMultiple Languages

IBM Granite-4.0-H-Tiny is a hybrid Mamba-2/Transformer model optimized for Apple silicon chips, using 3-bit quantization technology, and is designed for long context, efficient inference, and enterprise use. This model combines the Mamba-2 architecture with the Mixture of Experts (MoE) technology, significantly reducing memory usage while maintaining expressiveness.

Natural Language ProcessingMlxMultiple Languages

Qwen3-Next-80B-A3B-Instruct is an efficient sparse mixture of experts model developed by the Tongyi Qianwen team. It has a total of 80 billion parameters, with only 3 billion active parameters. The model adopts an innovative hybrid attention mechanism and a MoE architecture with an extremely low activation rate, significantly improving inference efficiency while maintaining strong performance. It natively supports a context length of 262,000 tokens and can be extended to 1 million tokens.

Natural Language ProcessingTransformers

An expert hybrid model based on the Llama3 architecture, which combines four 8B models to form a powerful model with 25B parameters and is good at creative writing and role - playing.

Natural Language ProcessingGgufEnglish