Open Source AI Model Research

Open Source Models NEMOTRON DEV DOCS \ Repo List of various open source LLMS instaclustr Meta / LLaMA family 1. Llama (various versions) - Models hub: <https://huggingface.co/meta-llama huggingface - Docs: model cards on each HF page. - High‑quality open‑weight LLM family widely used as a base for instruction, code, and chat models. github Mistral & Mixtral 2. Mistral / Mixtral - Models hub: <https://huggingface.co/mistralai huggingface - Docs: HF model cards and Mistral’s site linked from cards. - Strong open‑weight LLMs and MoE models, commonly used for efficient inference and fine‑tuning. instaclustr Qwen family 3. Qwen - Models hub: <https://huggingface.co/Qwen huggingface - Docs: model cards and linked GitHub docs per version. - Competitive multilingual/code‑capable LLMs from Alibaba with permissive licenses for many variants. instaclustr GLM family 4. GLM - Models hub: <https://huggingface.co/THUDM (GLM series) huggingface - Docs: HF cards + THUDM GitHub links. - Chinese‑origin LLMs with strong reasoning and coding performance, often recommended as top open‑source choices in 2026. kilo Falcon series 5. Falcon - Repo / info: <https://huggingface.co/tiiuae (Falcon models) github - Docs: HF model cards and TII blog docs linked there. github - Early high‑profile open‑weight LLM family known for solid general performance and permissive licensing. instaclustr MPT series 6. MPT - Repo/info: <https://huggingface.co/mosaicml (MPT models) github - Docs: HF cards + MosaicML docs linked from them. github - Decoder‑only LLMs trained on mixed code/text, including variants optimized for chat and instruction. github Gemma / Gemma‑like (Google) 7. Gemma - Models hub: <https://huggingface.co/google (Gemma models) huggingface - Docs: model cards + Google’s documentation linked there. - Google’s smaller open‑weight LLMs tuned for efficiency and on‑device/edge use. instaclustr Phi family 8. Phi - Models hub: “microsoft” org on HF (Phi‑class models): <https://huggingface.co/microsoft huggingface - Docs: model cards linking to Microsoft blog posts. - Small, efficient models emphasizing high quality per parameter, often used where VRAM is limited. instaclustr DeepSeek family 9. DeepSeek - Models hub: <https://huggingface.co/deepseek-ai github - Docs: model cards + DeepSeek documentation. - Open‑weight LLMs with strong reasoning/coding, some versions reported near GPT‑4‑class performance. github Nemotron / NVIDIA models 10. Nemotron - Models hub: <https://huggingface.co/nvidia (Nemotron, etc.) huggingface - Docs: HF cards + NVIDIA docs. - NVIDIA’s open‑weight LLMs used heavily for experimentation on their hardware stack. instaclustr Multimodal & vision‑language 11. LLaVA (vision + language) - Repo: <https://github.com/haotian-liu/LLaVA github - Docs: in repo README and linked docs. - Popular open VLM that combines a vision encoder with an LLM for image‑conditioned chat and reasoning. github 12. OpenCLIP - Repo: <https://github.com/mlfoundations/openclip github - Docs: README + examples. - Open implementation and checkpoints for CLIP‑style vision‑language models used in retrieval and as components in generative pipelines. github 13. Unlimited‑OCR (Baidu VLM) - Repo: <https://github.com/baidu/Unlimited-OCR youtube - Docs: in repo. - Open‑weights vision‑language model for fast OCR and document understanding. youtube Diffusion / image & media generation 14. Stable Diffusion & variants - Base models hub: e.g. <https://huggingface.co/stabilityai huggingface - Docs: model cards and Stability AI docs. - Open diffusion models for image generation, fine‑tuning, and downstream creative tasks. bentoml 15. Open‑source video generation models - Example framework (video editing / generation): Hyperframes repo: <https://github.com/heygen-com/hyperframes youtube - Docs: README and linked docs. - AI‑native video editing and generation pipeline, representative of open media‑generation stacks. youtube Curated lists & indexes (discover “all” models) 16. Hugging Face Models Index - Site: <https://huggingface.co/models huggingface - Docs: platform docs at <https://huggingface.co/docs huggingface - Main index of 2M+ models (LLMs, vision, audio, diffusion); primary way to find specific open‑weight models and read their docs. huggingface 17. Open LLMs for commercial use (curated) - Repo: <https://github.com/eugeneyan/open-llms github - Docs: in repo. - List of LLMs with commercial‑friendly licenses plus links to their model cards/repos. github 18. Awesome‑LLM - Repo: <https://github.com/hannibal046/awesome-llm github - Docs: curated README. - Curated list of LLM papers, tools, and publicly available checkpoints/APIs, including pointers to many model families. github 19. Top open‑source LLMs 2026 overviews - Example: “Top 7 open source LLMs for 2026” article: <https://www.instaclustr.com/education/open-source-ai/top-7-open-source-llms-for-2026/ instaclustr - Docs: article text. - High‑level comparison of leading open‑source LLM families with links out…