This model is the 8-billion parameter version in the Llama 3.1 series released by Meta. It has been processed and post-processed by the AMD Quark quantization tool, optimizing the deployment and inference efficiency on specific hardware. It is a multilingual text generation model that supports multiple languages and performs excellently in benchmark tests such as MMLU.

Natural Language ProcessingOnnxMultiple Languages

This is a GGUF quantized version converted from the Qwen3-4B-Instruct model, using Q8 quantization technology, suitable for the llama.cpp inference framework. The model is optimized through model fusion technology and performs excellently in academic benchmark tests such as MMLU and GPQA.

Natural Language ProcessingTransformersEnglish

Typhoon T1 3B is the first product in the new open reasoning model series developed by SCB 10X. Based on the Llama 3.2 architecture, this model has cross-domain reasoning capabilities and performs excellently in benchmark tests such as GPQA, MMLU Pro, and AI Mathematical Olympiad. It especially supports the generation of Thai reasoning processes.

Natural Language ProcessingTransformersMultiple Languages

Mixtral MOE 2x10.7B is a large language model based on the Mixture of Experts architecture, which combines the advantages of two base models, Sakura-SOLAR-Instruct and CarbonVillain. This model performs excellently in text generation tasks and has been evaluated on multiple public datasets, including benchmark tests such as AI2 Reasoning Challenge, HellaSwag, and MMLU.

Natural Language ProcessingTransformers