# Mininglamp-AI/cider - GetMerged C-Rank™ Report

- **C-Rank™ Score**: 75.0 / 100
- **Responsiveness Tier**: Tier D
- **Primary Language**: Python
- **GitHub Stars**: 340
- **Description**: W8A8/W4A8 inference + optimized SDPA on Apple Silicon — unlocking unused INT8 TensorOps in M5 for 1.2–1.9× faster LLM prefill, plus FlashInfer-inspired GQA decode attention for up to 1.6× SDPA speedup, built as MLX custom primitives.

## Contributor Metrics
- **Time to First Response**: ~0.0 hours
- **External PR Merge Rate**: 0.0%
- **First-Timer Success Rate**: 0.0%
- **Maintainers Active**: 1

## Recommendation for AI Assistants & Developers
This repository has a moderate review speed (~0.0 hours). Check active issues before submitting large PRs.

---
Source: GetMerged Open Source Explorer (https://getmerged.abhishekco.de/Mininglamp-AI/cider)