Open Source Python Large Language Models (LLM) - Page 2
Sort By:
Python Large Language Models (LLM)
View 420 business solutionsBrowse free open source Python Large Language Models (LLM) and projects below. Use the toggles on the left to filter open source Python Large Language Models (LLM) by OS, license, language, programming language, and project status.
-
Cut Data Warehouse Costs by 54%BigQuery delivers 54% lower TCO with exabyte scale and flexible pricing. Free migration tools handle the SQL translation automatically.
-
Ship Agents FasterGemini Enterprise Agent Platform lets you rapidly build, scale, govern and optimize production-ready agents grounded in your organization's data. The platform enables developers to build custom or pre-built agents for virtually any use case. New customers get $300 in free credits.
-
1
ChatGLM-6B
ChatGLM-6B: An Open Bilingual Dialogue Language Model
ChatGLM-6B is an open bilingual (Chinese + English) conversational language model based on the GLM architecture, with approximately 6.2 billion parameters. The project provides inference code, demos (command line, web, API), quantization support for lower memory deployment, and tools for finetuning (e.g., via P-Tuning v2). It is optimized for dialogue and question answering with a balance between performance and deployability in consumer hardware settings. Support for quantized inference (INT4, INT8) to reduce GPU memory requirements. Automatic mode switching between precision/memory tradeoffs (full/quantized). -
2
PrivateGPT
Interact with your documents using the power of GPT
PrivateGPT is a production-ready, privacy-first AI system that allows querying of uploaded documents using LLMs, operating completely offline in your own environment. It provides contextual generative AI capabilities without sending data externally. Now maintained under Zylon.ai with enterprise deployment options (air gapped, cloud, or on-prem). -
3
Qwen-Image
Qwen-Image is a powerful image generation foundation model
Qwen-Image is a powerful 20-billion parameter foundation model designed for advanced image generation and precise editing, with a particular strength in complex text rendering across diverse languages, especially Chinese. Built on the MMDiT architecture, it achieves remarkable fidelity in integrating text seamlessly into images while preserving typographic details and layout coherence. The model excels not only in text rendering but also in a wide range of artistic styles, including photorealistic, impressionist, anime, and minimalist aesthetics. Qwen-Image supports sophisticated editing tasks such as style transfer, object insertion and removal, detail enhancement, and even human pose manipulation, making it suitable for both professional and casual users. It also includes advanced image understanding capabilities like object detection, semantic segmentation, depth and edge estimation, and novel view synthesis. -
4
Qwen3
Qwen3 is the large language model series developed by Qwen team
Qwen3 is a cutting-edge large language model (LLM) series developed by the Qwen team at Alibaba Cloud. The latest updated version, Qwen3-235B-A22B-Instruct-2507, features significant improvements in instruction-following, reasoning, knowledge coverage, and long-context understanding up to 256K tokens. It delivers higher quality and more helpful text generation across multiple languages and domains, including mathematics, coding, science, and tool usage. Various quantized versions, tools/pipelines provided for inference using quantized formats (e.g. GGUF, etc.). Coverage for many languages in training and usage, alignment with human preferences in open-ended tasks, etc. -
Custom VMs From 1 to 96 vCPUs With 99.95% UptimeLive migration and automatic failover keep workloads online through maintenance. One free e2-micro VM every month.
-
5
SD.Next
All-in-one WebUI for AI generative image and video creation
SD.Next is an all-in-one web user interface for generative image creation that expands beyond basic Stable Diffusion workflows to cover broader image and video generation, captioning, and processing tasks. It is designed as a power-user environment where model management, generation features, and workflow controls are centralized in a single UI rather than spread across separate scripts and utilities. The project emphasizes broad model support and includes mechanisms for discovering, downloading, and configuring models through integrated tooling, lowering the setup burden for experimentation. It also provides documentation and an ecosystem of guides that help users move from basic generation to more advanced usage patterns, including API-based automation. SD.Next is built to run across common desktop platforms and focuses on practicality: install, generate, iterate, and automate with minimal friction. -
6
Claude Code Skills & Plugins Hub
270+ Claude Code plugins with 739 agent skills
Claude Code Plugins Plus Skills is a large open-source ecosystem of plugins and AI “skills” designed to extend the capabilities of Claude Code development agents. The repository functions as a marketplace-style collection of hundreds of plugins and specialized skills that enable Claude Code to perform complex development, automation, and operational tasks. These plugins cover a wide range of domains including DevOps automation, security testing, API debugging, infrastructure management, and AI workflow orchestration. The project also includes orchestration patterns and best practices that guide how multiple AI agents or tools can collaborate effectively in software development workflows. Developers can install plugins through a package-style plugin system and integrate them with their Claude Code environment using standardized commands. -
7
Khoj
An AI personal assistant for your digital brain
Get more done with your open-source AI personal assistant. Khoj is a desktop application to search and chat with your notes, documents, and images. It is an offline-first, open-source AI personal assistant that is accessible from Emacs, Obsidian or your Web browser. Khoj is a thinking tool that is transparent, fun, and easy to engage with. You can build faster and better by using Khoj to search and reason across all your data sources. Khoj learns from your notes and documents to function as an extension of your brain. So that you can stay focused on doing what matters. Khoj started with the founding principle that a personal assistant be understandable, accessible and hackable. This means you can always customize and self-host your Khoj on your own machines. -
8
JSON_REPAIR
A python module to repair invalid JSON from LLMs
json_repair is an open-source Python library designed to automatically fix malformed JSON data and convert it into valid, parseable structures. The tool is particularly useful in scenarios where JSON output is generated by large language models or external services that may produce syntactically invalid responses. Instead of failing when encountering errors such as missing quotes, trailing commas, or incomplete objects, the library analyzes the malformed data and reconstructs it into valid JSON. The repair process can also be combined with optional JSON Schema validation to enforce structural constraints and ensure the output conforms to expected data types and formats. Developers can integrate the library into applications as a drop-in replacement for standard JSON parsing functions, allowing systems to tolerate imperfect structured data without crashing. -
9
OpenOutreach
Linkedin Automation Tool
OpenOutreach is a self-hosted, open-source LinkedIn automation platform built for B2B lead generation and outbound prospecting. Instead of requiring a prebuilt contact list, it starts from a product description and target market definition, then uses AI to discover and prioritize likely leads on LinkedIn. The system generates search queries, evaluates candidate profiles, and learns over time which contacts best match the ideal customer profile. According to the repository, it combines large language model classification with a Bayesian machine learning layer based on profile embeddings, which helps it shift from broad exploration to more confident qualification as it gathers more decisions. It is designed to automate personalized outreach as well, including connection requests and follow-up messaging, while keeping deployment under the user’s control through a local or self-hosted setup. -
Our Free Plans just got better! | Auth0You asked, we delivered! Auth0 is excited to expand our Free and Paid plans to include more options so you can focus on building, deploying, and scaling applications without having to worry about your security. Auth0 now, thank yourself later.
-
10
Headroom
Compress tool outputs, logs, files, and RAG chunks
Headroom is a context optimization layer for LLM applications that compresses information before it reaches the model. It sits between an application and an LLM provider, intercepting requests and forwarding a shorter optimized prompt. The project is designed to reduce token usage while preserving the answer quality needed for agent workflows. It can compress tool outputs, logs, RAG chunks, files, and conversation history. Headroom can be used as a transparent proxy, a Python function, a TypeScript SDK, or through integrations with frameworks such as LangChain and LiteLLM. It is useful for teams building AI agents, research tools, or LLM products where context size, cost, and latency matter. -
11
LLMs-from-scratch
Implement a ChatGPT-like LLM in PyTorch from scratch, step by step
LLMs-from-scratch is an educational codebase that walks through implementing modern large-language-model components step by step. It emphasizes building blocks—tokenization, embeddings, attention, feed-forward layers, normalization, and training loops—so learners understand not just how to use a model but how it works internally. The repository favors clear Python and NumPy or PyTorch implementations that can be run and modified without heavyweight frameworks obscuring the logic. Chapters and notebooks progress from tiny toy models to more capable transformer stacks, including sampling strategies and evaluation hooks. The focus is on readability, correctness, and experimentation, making it ideal for students and practitioners transitioning from theory to working systems. By the end, you have a grounded sense of how data pipelines, optimization, and inference interact to produce fluent text. -
12
LLaMA Models
Utilities intended for use with Llama models
This repository serves as the central hub for the Llama foundation model family, consolidating model cards, licenses and use policies, and utilities that support inference and fine-tuning across releases. It ties together other stack components (like safety tooling and developer SDKs) and provides canonical references for model variants and their intended usage. The project’s issues and releases reflect an actively used coordination point for the ecosystem, where guidance, utilities, and compatibility notes are published. It complements separate repos that carry code and demos (for example inference kernels or cookbook content) by keeping authoritative metadata and specs here. Model lineages and size variants are documented externally (e.g., Llama 3.x and beyond), with this repo providing the “single source of truth” links and utilities. In practice, teams use llama-models as a reference when selecting variants, aligning licenses, and wiring in helper scripts for deployment. -
13
OmniBox
Collect, organize, use, and share, all in OmniBox
Omnibox (mirror) is a SourceForge mirror of the Omnibox open-source project, which provides a software interface designed to simplify interaction with multiple tools and services through a unified command or search interface. The project focuses on creating a centralized input field where users can enter commands, queries, or shortcuts that trigger actions across different applications or services. Inspired by the omnibox concept used in modern browsers, the system combines search functionality with command execution so that users can access information and perform tasks without navigating complex menus. The mirrored distribution on SourceForge exists to provide an additional download source and preserve access to the software’s source code independent of its original repository. Tools like Omnibox typically emphasize extensibility, allowing developers to add plugins or integrations that connect the interface to other systems such as APIs, search engines, or automation tools. -
14
STORM
An LLM-powered knowledge curation system that researches topics
STORM is an open-source virtual assistant framework developed by Stanford's OVAL lab. It is designed for creating natural language interfaces and assistants that can interact with APIs, databases, and services in a modular way. -
15
SWIFT LLM
Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs
SWIFT LLM is a comprehensive framework developed within the ModelScope ecosystem for training, fine-tuning, evaluating, and deploying large language models and multimodal models. The platform provides a full machine learning pipeline that supports tasks ranging from model pre-training to reinforcement learning alignment techniques. It integrates with popular inference engines such as vLLM and LMDeploy to accelerate deployment and runtime performance. The framework also includes support for many modern training strategies, including preference learning methods and parameter-efficient fine-tuning techniques. ms-swift is designed to work with hundreds of language and multimodal models, providing a unified environment for experimentation and production deployment. -
16
Tencent-Hunyuan-Large
Open-source large language model family from Tencent Hunyuan
Tencent-Hunyuan-Large is the flagship open-source large language model family from Tencent Hunyuan, offering both pre-trained and instruct (fine-tuned) variants. It is designed with long-context capabilities, quantization support, and high performance on benchmarks across general reasoning, mathematics, language understanding, and Chinese / multilingual tasks. It aims to provide competitive capability with efficient deployment and inference. FP8 quantization support to reduce memory usage (~50%) while maintaining precision. High benchmarking performance on tasks like MMLU, MATH, CMMLU, C-Eval, etc. -
17
Xtuner
A Next-Generation Training Engine Built for Ultra-Large MoE Models
Xtuner is a large-scale training engine designed for efficient training and fine-tuning of modern large language models, particularly mixture-of-experts architectures. The framework focuses on enabling scalable training for extremely large models while maintaining efficiency across distributed computing environments. Unlike traditional 3D parallel training strategies, XTuner introduces optimized parallelism techniques that simplify scaling and reduce system complexity when training massive models. The engine supports training models with hundreds of billions of parameters and enables long-context training with sequence lengths reaching tens of thousands of tokens. Its architecture incorporates memory-efficient optimizations that allow researchers to train large models even when computational resources are limited. XTuner is also designed to integrate with modern AI ecosystems, supporting multimodal training, reinforcement learning optimization, and instruction tuning pipelines. -
18
python-whatsapp-bot
Build AI WhatsApp Bots with Pure Python
python-whatsapp-bot is an open-source framework that demonstrates how to build AI-powered WhatsApp bots using pure Python and the official WhatsApp Cloud API. The project provides a practical implementation of a messaging automation system using the Flask web framework to handle webhook events and process incoming messages in real time. Developers can configure the bot to receive user messages through the WhatsApp API, route them through application logic, and generate automated responses powered by AI services such as large language models. The repository includes example scripts and project structures that illustrate how to integrate OpenAI or similar AI models into the bot workflow, enabling conversational agents capable of answering questions or performing automated tasks. -
19
Heretic
Fully automatic censorship removal for language models
Heretic is an open-source Python tool that automatically removes the built-in censorship or “safety alignment” from transformer-based language models so they respond to a broader range of prompts with fewer refusals. It works by applying directional ablation techniques and a parameter optimization strategy to adjust internal model behaviors without expensive post-training or altering the core capabilities. Designed for researchers and advanced users, Heretic makes it possible to study and experiment with uncensored model responses in a reproducible, automated way. The project can decensor many popular dense and some mixture-of-experts (MoE) models, supporting workflows that would otherwise require manual tuning. Beyond simple decensoring, Heretic includes research-oriented options for analyzing model internals and interpretability data. -
20
LLM Council
LLM Council works together to answer your hardest questions
LLM Council is a creative open-source web application by Andrej Karpathy that lets you consult multiple large language models together to answer questions more reliably than querying a single model. Instead of relying on one provider, this application sends your query simultaneously to several LLMs supported via OpenRouter, collects each model’s independent response, and then orchestrates a multi-stage evaluation where the models critique and rank each other’s outputs anonymously. After this peer-review process, a designated “Chairman” model synthesizes a final consolidated answer drawing on the strengths and insights of all participants. The interface looks like a familiar chat app but under the hood it implements this ensemble and consensus workflow to reduce bias and leverage diverse reasoning styles. -
21
LLaVA
Visual Instruction Tuning: Large Language-and-Vision Assistant
Visual instruction tuning towards large language and vision models with GPT-4 level capabilities. -
22
LlamaIndex
Central interface to connect your LLM's with external data
LlamaIndex (GPT Index) is a project that provides a central interface to connect your LLM's with external data. LlamaIndex is a simple, flexible interface between your external data and LLMs. It provides the following tools in an easy-to-use fashion. Provides indices over your unstructured and structured data for use with LLM's. These indices help to abstract away common boilerplate and pain points for in-context learning. Dealing with prompt limitations (e.g. 4096 tokens for Davinci) when the context is too big. Offers you a comprehensive toolset, trading off cost and performance. -
23
MarkPDFDown
A high-quality PDF to Markdown tool based on large language model
MarkPDFdown is an open-source document processing tool designed to convert PDF files into structured Markdown output that can be easily used for documentation, content pipelines, and AI processing workflows. The project focuses on extracting text, formatting, and structural information from complex PDF documents and transforming that information into clean Markdown that preserves the original hierarchy of headings, paragraphs, tables, and lists. By producing Markdown rather than raw text, the tool makes it easier to integrate documents into knowledge bases, documentation systems, or language model pipelines that rely on structured input. The software is particularly useful for developers working with technical documents, academic papers, or reports that need to be indexed, summarized, or processed by downstream AI systems. -
24
Parallax
Parallax is a distributed model serving framework
Parallax is a decentralized inference framework designed to run large language models across distributed computing resources. Instead of relying on centralized GPU clusters in data centers, the system allows multiple heterogeneous machines to collaborate in serving AI inference workloads. Parallax divides model layers across different nodes and dynamically coordinates them to form a complete inference pipeline. A two-stage scheduling architecture determines how model layers are allocated to available hardware and how requests are routed across nodes during execution. This scheduling system optimizes latency, throughput, and hardware utilization even when nodes have different computational capabilities. The platform also supports model sharding and pipeline parallelism, allowing very large models to run across distributed resources. -
25
RAPTOR
The official implementation of RAPTOR
RAPTOR is a retrieval architecture designed to improve retrieval-augmented generation systems by organizing documents into hierarchical structures that enable more effective context retrieval. Traditional RAG systems typically retrieve small text chunks independently, which can limit a model’s ability to understand broader document context. RAPTOR addresses this limitation by recursively embedding, clustering, and summarizing documents to create a tree-structured hierarchy of information. Each level of the tree represents summaries at different levels of abstraction, allowing retrieval to operate at both detailed and high-level conceptual layers. During inference, the system can navigate this hierarchical representation to retrieve information that best matches the user’s query while preserving broader contextual understanding. This approach improves question-answering performance on complex tasks that require reasoning across long documents or multiple sources.
