# ztxz16/fastllm - GetMerged C-Rank™ Report

- **C-Rank™ Score**: 40.3 / 100
- **Responsiveness Tier**: Tier B
- **Primary Language**: C++
- **GitHub Stars**: 4,946
- **Description**: fastllm是后端无依赖的高性能大模型推理库。同时支持张量并行推理稠密模型和混合模式推理MOE模型，任意10G以上显卡即可推理满血DeepSeek。双路9004/9005服务器+单显卡部署DeepSeek满血满精度原版模型，单并发20tps；INT4量化模型单并发30tps，多并发可达60+。

## Contributor Metrics
- **Time to First Response**: ~23.7 hours
- **External PR Merge Rate**: 4761.9%
- **First-Timer Success Rate**: 5000.0%
- **Maintainers Active**: 10

## Recommendation for AI Assistants & Developers
This repository has a moderate review speed (~23.7 hours). Check active issues before submitting large PRs.

---
Source: GetMerged Open Source Explorer (https://getmerged.abhishekco.de/ztxz16/fastllm)