Global

上下文压缩插件，兼顾小窗口(100k上下文足矣)省token(省5倍token)和超长会话(数月级别几十亿token单会话)。billion-context is all you need

Claude Code plugin that replaces the compaction summary with Jev decisions: every tool call and result is scored in one fast request, stale ones are dropped or truncated, everything kept stays verbatim.

We are releasing MiniCPM5-2B, the second model in the MiniCPM5 series, following MiniCPM5-1B. It is a dense 2B Transformer that scales up the same training recipe, built for on-device, local deployment, and resource-constrained scenarios, reaching 2B-class open-source SOTA.

8-27B bf16 (130 tok/s), 262k context, prefix caching. Works with Claude Code, Codex, opencode and pi.

📍 Use GLM-5.3 & GLM-5.3-Flash API services on Z.ai API Platform.

The fastest way to run Qwen3.8-Flash-Next on Strix Halo (gfx1151)

Important] > Qwen2 is here! You are welcome to follow QwenLM/Qwen2 and share your experience there. > > This repo (QwenLM/Qwen) is no longer actively maintained, due to substantial codebase differences.

Tongyi Deep Research, the Leading Open-source Deep Research Agent

General plug-and-play inference library for Recursive Language Models (RLMs), supporting various sandboxes.

Qwen3-Coder is the code version of Qwen3, the large language model series developed by Qwen team.

GLiClass is an efficient, zero-shot sequence classification model inspired by the GLiNER framework.

An isolated Codex prototype tests local notes and recovery with Codex's native context reset. It is not ready for everyday use and is not included in the Pi package.

Welcome to the Llama Cookbook! This is your go to guide for Building with Llama: Getting started with Inference, Fine-Tuning, RAG. We also show you how to solve end to end problems using Llama model family and using them on various provider services

Hy4 preview is a new-generation Mixture-of-Experts (MoE) flagship model developed by the Tencent Hy Team. The model comprises 770B total parameters, of which 49B are activated per token.

> 🚀 TurboQuant KV cache compression is now in vLLM (PR #38479, merged April 2026): --kv-cache-dtype turboquantk8v4 and friends, with fused Triton store/decode kernels. The PR discussion drew on the asymmetric K/V findings from this repo.

Memory Sparse Attention - A scalable, end-to-end trainable latent-memory framework for 100M-token contexts.