Top 14 Open-Source startups

Updated: Dec 16, 2025
|
These startups develop open-source ML frameworks, LLMs and other AI models.
1
Domyn
Country: Italy | Funding: €40M
iGenius develops enterprise AI platform that allows to create and manage AI applications and agents via centralized control system. iGenius AI agents operate in a companie's closed loop and adapt to specific business context. iGenius is also developing its own enterprise LLM model with an open enterprise license. Domyn models are designed for industries with strict compliance standards, such as financial services, public sector and heavy industry. They are trained in the FP8 format, optimize the use of computing resources, accelerate computation and reduce costs and energy consumption. Continuous training on company's own data allows the language model to incorporate domain-specific knowledge and adapt to the organization's requirements. The architecture of Domyn language models is designed to meet the highest security and compliance standards in regulated sectors.
2
Databricks
Country: USA | Funding: $25.8B
Databricks is developing cloud-based platform Databricks Data Intelligence for working with Big Data and machine learning. It brings together everything needed for the full data cycle: storage, processing, analysis, model creation, training and deployment. It includes MLflow - an open-source system for managing machine learning models, experiment tracking and automated deployment. The platform provides interactive notebooks (like Jupyter) for writing code in Python, SQL, R and Scala and running it on clusters. Databricks allows to develop your own models, AI agents and generative AI applications based on company's own data.
3
DeepSeek
Country: China
DeepSeek is specializing in the development of large-scale open-source language models. The company unexpectedly grew out of a Chinese hedge fund and disrupted the AI ​​market with its technologies, which significantly reduced the cost of training and inference of LLM models. These technologies include Mixture of Experts - a fragmented model splitting into individual "experts" so that only a small portion of the model could be used for each task. Unlike traditional language models, which rely heavily on supervised fine-tuning, DeepSeek relies primarily on reinforcement learning. It also came up with the idea of ​​using distilled data from OpenAI models for training. DeepSeek provides its online chatbot and an API for business. Its responses are often criticized due to Chinese censorship.
4
Numenta
Country: USA
Numenta is a startup by Jeff Hawkins, the author of renowned books on AI ("On Intelligence" and "A Thousand Brains: A New Theory of Intelligence") and his own theory of brain structure. Therefore, the startup's primary goal was initially to understand the neocortex (the primary cerebral cortex) and apply its principles to the creation of machine intelligence. The theory, called Hierarchical Temporal Memory (HTM) describes a learning and prediction model inspired by the cerebral cortex. The company has published open-source code of its framework (Thousand Brains Project) - a platform for sensory-motor learning. The company's main product, the Numenta Platform for Intelligent Computing (NuPIC), is also based on algorithms, data structures and architecture inspired by the brain.
5
Hugging Face
Country: France | Funding: $395.2M
Hugging Face is a platform where the machine learning community collaborates on models, datasets and applications. It contains a vast catalog of open-source ML models, frameworks, inference providers, datasets, libraries and apps. It also allows users to combine machine learning models and datasets and create collaborative projects/Spaces for developing AI systems. An enterprise version of this workspace is available for on-premises deployment. The platform hosts a community that provides opportunities for networking, knowledge sharing and building own ML portfolio and reputation. Hugging Face also develops own ML tools, including the Transformers library, designed for natural language processing applications.
6
Mistral AI
Country: France | Funding: €3B
Mistral is a leading European AI company that develops own LLM models and provides AI products and solutions for the enterprise. The startup competes with American AI giants by focusing on the European market and local specifics/languages. Furthermore, Mistral is committed to open source - nearly all of its models are open. The company has created the online assistant Le Chat, which is available on the web and mobile app, free and by subscription. It also offers businesses the AI ​​Studio platform for creating and managing the full lifecycle of enterprise models and agents, Mistral Code - AI-powered software development system and Mistral Compute - European AI cloud that combines high-performance GPUs, ready-to-use orchestration and advanced model building tools into a single, fully integrated platform.
7
Reflection AI
Country: USA | Funding: $2.1B
Reflection AI, originally created as a coding agent platform, now positions itself as an open-source alternative to leading AI companies like OpenAI and Anthropic, and as a Western counterpart to Chinese AI companies like DeepSeek. It is building a large-scale LLM and reinforcement learning platform capable of training massive mixed-expert (MoE) models - the technology that has enabled DeepSeek to significantly reduce training and inference costs. Reflection AI will publish model weights -the fundamental parameters that determine an AI system's performance, for public use, while the datasets and all training processes will remain proprietary. The company aims to generate revenue from large enterprises building products based on Reflection AI models, and from governments developing "sovereign AI systems".
8
Stability AI
Country: UK | Funding: $256M
Stability AI is developing an open-source model for generating images from text descriptions - Stable Diffusion. The model enables creation of a wide range of image styles, including 3D, photography, painting, line art and more. The model can be installed on a private server or accessed via Stability AI API. The company also provides a Studio for editing images, transforming styles, up-scaling and improving image quality. In addition to the image generation model, Stability AI also develops text-to-video, -audio and -3D models and is collaborating with ARM to optimize its models for mobile devices. Stability AI is constantly facing legal action from stock photo companies like Getty Images for misusing their collection to train their AI-image generation system.
9
LangChain
Country: USA | Funding: $160M
LangChain is a platform providing open-source frameworks and tools for developing and deploying agents based on language models. LangChain began as a project developing open-source add-on tools for creating language models (web search, API calls and database interactions). As modern model developers expanded their infrastructure, LangChain evolved into a platform for creating enterprise agents. The company's core products include the LangChain agent builder, LangGraph - orchestration and context/memory management tool, and LangSmith - testing/observation tool. LangChain remains extremely popular among open-source developers.
10
SingularityNET
Country: Netherlands | Funding: $66M
SingularityNET is attempting to create Artificial General Intelligence (AGI) in format of the open, decentralized blockchain platform that hosts multiple AI services and algorithms. Developers can publish their services on the SingularityNET network, where they are available for anyone to use, and can charge for their services using the AGIX token. The company has also created an AI-DSL (domain-specific language), which will allow individual, highly specialized services to form collaborative workflows and perform more complex tasks than any single service could, as well as its own open-source AGI framework, OpenCog Hyperon.
11
Genmo
Country: USA | Funding: $58.4M
Genmo is an artificial intelligence creative content generation platform that specializes in developing creative products.
12
HyperWrite
Country: USA | Funding: $5.4M
AI-powered open-source writing assistant with web search and citations
13
All Hands AI
Country: USA | Funding: $5M
Open Source Agents for Developers. Use AI to tackle the toil in your backlog, so you can focus on what matters: hard problems, creative challenges, and over-engineering your dotfiles
14
01.AI
Country: China
01.AI is developing a language model for the Chinese market that enhances productivity, creating significant economic and societal value.
  See also:
Siddhant Patel
Editor: Siddhant Patel
Siddhant Patel is a senior editor for AI-Startups. He is based out of India and has previously worked at publications including Huffington Post and The Next Web. Siddhant has a special interest in artificial intelligence and has spent a decade covering the rapidly-evolving business and technology of the industry. Siddhant graduated from the Indian Institute of Science (Bengaluru). When he’s not writing, Siddhant is also a developer and has a deep historical knowledge of the computer industry for the past 50 years. You can contact Siddhant at sidpatel(at)ai-startups(dot)pro