---
title: README
emoji: ⚡
colorFrom: yellow
colorTo: gray
sdk: static
pinned: false
---

Linear-Next: The Definitive Linear Attention Model Leaderboard

Comprehensive benchmarks and rankings for the latest AI models. Compare performance across key metrics to find the perfect model for your needs.

- Training Framework: [flame](https://github.com/Linear-Next/flame)
- Training Corpus: General, Code, Math, Reasoning
- Training Budget: Small, Medium, Large
- Model Size: 90M, 410M, 1B, 3B, 7B
- Model Variant: Linear Attention, Sparse Attention, Hybrid
- Benchmarks: Knowledge QA, General Reasoning, Math, Code, Long Context

GitHub: https://github.com/Linear-Next