# High-precision reasoning

Based on andrewzh's Absolute_Zero_Reasoner-Coder-14b model, this is a version quantized using llama.cpp's imatrix, suitable for reasoning and code generation tasks.

This is a 15B-parameter large language model built by the ServiceNow Language Model (SLAM) lab, quantized using the llama.cpp tool, suitable for local inference deployment.

A Mishima Imatrix quantized version based on Qwen3-4B, enhanced with specific datasets for prose-style generation

phi-4 is an open-source language model developed by Microsoft Research, focusing on high-quality data and reasoning capabilities, suitable for memory/computation-constrained environments.

A 72B-parameter multimodal foundation model released by ByteDance Research, specializing in image-text-to-text tasks

GLM-4-9B is an open-source version of the latest generation of pre-trained models in the GLM-4 series launched by Zhipu AI. It performs excellently in the evaluation of datasets such as semantics, mathematics, reasoning, code, and knowledge, and has advanced features such as multilingual support.

A model post-trained based on Meta-Llama-3-70B, specifically optimized for role-playing tasks while excelling in general domain instruction following and reasoning tasks.

YiSM-34B-0rn is a large language model based on the fusion of Yi-1.5-34B and Yi-1.5-34B-Chat, designed to balance instruction-following capabilities with foundational model characteristics.

InternLM2-20B is the second-generation model with 20B parameters, supporting 200k-word ultra-long context, excelling in reasoning, mathematics, and programming capabilities.

A 7B-parameter large language model based on the Mistral architecture, utilizing SLERP fusion technology to merge weights from multiple high-performance models