Natural Language Processing

MLC-LLM is a free open-source universal LLM deployment engine that runs models on any device — phones, browsers, gaming consoles, GPUs. Apache 2.0, powered by TVM/Apache. Best free framework for cross-platform AI deployment.

TensorRT-LLM by NVIDIA is the free open-source library for ultra-fast LLM inference on NVIDIA GPUs. Apache 2.0, supports all major LLMs, delivers 2-4x speedup vs vLLM. Best free framework for production NVIDIA AI deployment.

Poro 34B by SiloGen and the University of Turku is a free open-source 34B bilingual Finnish-English LLM. Apache 2.0, trained on 1 trillion tokens. Best free LLM for Finnish, Nordic, and other European low-resource languages.

Mamba-2.8B is a free open-source state-space model that beats transformers of the same size with 5x faster inference and unlimited context length. Apache 2.0, perfect for long-document tasks and edge AI.

Yi-34B by 01.AI is a free open-source 34-billion-parameter bilingual LLM with 200K context window. Strong English & Chinese performance, Apache 2.0 license, beats Llama 2-70B on many benchmarks. Best mid-size free LLM.

StableLM 3.5 by Stability AI is a free 3-billion-parameter compact LLM optimized for fast on-device inference. Strong multilingual support, runs on laptop CPU. Perfect for indie developers building local AI assistants.

Vicuna-13B v1.5 is a free open-source chat AI fine-tuned from Llama 2 on 125K ShareGPT conversations. Reaches 90% of ChatGPT quality on benchmarks, runs on a single consumer GPU. Ideal for privacy-first chatbot deployments.

OLMo 1.7 by AI2 (Allen Institute) is the most transparent open-source LLM ever — full training code, data (Dolma), checkpoints, and weights all released. Apache 2.0, perfect for AI research and reproducibility.