A high-fidelity benchmark for AI Site Reliability Engineering agents with real-world failure scenarios.

Benchmark local LLMs for speed, memory, and quality

The official repository for Qwen series models, including new coding and collaborative updates.

Experimental userspace to run macOS binaries on Linux ARM

A tool for managing LLM-based agents and prompt engineering workflows.

Open-source, ad-free Android companion client for self-hosted LLM servers like Ollama and LocalAI.

A systematic framework for technical documentation focusing on four distinct content types.

A CLI tool to turn Instagram reels into a local, searchable Markdown knowledge base for LLMs.

A suite of Cinematic Language operators for evaluating text-to-video and reference-to-video generation.

A global multimodal benchmark and dataset for mathematical reasoning and retrieval.

Neural retrieval system for semantic code-to-code discovery across MediaWiki repositories.

Benchmark and evaluation harness for measuring operational stealth in AI security agents.

An automated red-teaming agent using self-play to discover prompt injection attacks.

An open-source serial console server based on Raspberry Pi Pico.

An open-source model distilled from DeepSeek focusing on reducing censorship.

A public database for verifying the lineage, licensing, and scan status of open AI models.

A self-hosted game library manager with save sync and server-side patching

A pipeline for identifying and using strategy-specific features to steer reasoning in LRMs.

A GPU-accelerated multi-agent driving simulator for perspective-view self-play.

A decoder-only any-to-any multimodal model supporting arbitrary inputs and outputs.