---
title: O
emoji: 🌍
colorFrom: blue
colorTo: green
sdk: static
pinned: false
---

# [OLM (Open Language Model)](https://openlm.io/)

[OLM (OpenLM RevShare Token)](https://openlm.io/) is the first-ever tokenized AI model, fully integrated on blockchain, made possible through the world's first Initial Model Offering (IMO). By merging AI development with blockchain technology, OLM establishes a decentralized ecosystem where AI models aren't just trained and deployed, but also offer inference services directly on the blockchain, allowing $OLM token holders to benefit from transparent revenue-sharing mechanisms.

# Key Focus Areas

- AI R&D: OLM is at the forefront of developing, training, and fine-tuning AI models, constantly pushing the boundaries of artificial intelligence research.
- Crypto AI Integration: OLM models power ORA's AI Oracle, bringing decentralized AI inference capabilities to blockchain applications.
- $OLM Tokenization: Revenue generated within the OLM ecosystem is shared transparently among $OLM token holders, aligning incentives across stakeholders.

# Models

- [OpenLM-Chat 1B](https://huggingface.co/OLMResearch/olm-chat-1b): A conversational model with 1 billion parameters, fine-tuned for natural language tasks to provide highly effective and engaging interactions in chat applications.
- [OpenLM-Chat 7B](https://huggingface.co/OLMResearch/olm-chat-7b): With 7 billion parameters, this model offers superior performance in complex conversational AI, designed for nuanced and intricate dialogue scenarios.
- [OpenLM-Score](https://huggingface.co/OLMResearch/olm-score-7b): A specialized evaluation model used in decentralized applications, such as f(A.I.)r Launch, ensuring fair assessment and transparent processes.

# Researches

OLM's research encompasses ongoing community-driven governance proposals aimed at enhancing the capabilities of decentralized AI. Key proposals include:
- [R&D of OLM Instruction Fine-Tuning Models](https://snapshot.org/#/olm.eth/proposal/0x74a697d9c2d90f9fd6484095a74d098c3923a84751d93b3b92dcc6ceb387bcc5): fine-tune OpenLM model with supervised instruction-tuning from human feedback (RLHF).
- [R&D of SearchOLM in ChatOLM](https://snapshot.org/#/olm.eth/proposal/0x8c41df88d8296e0cdb7a49d019d9b6990100a6e7e1f5eb3eefcc0dacdade9067): support imperative user requests and a chat interface with OpenLM model.

# Join the Decentralized AI Movement

Explore OLM's models and participate in building the future of AI on blockchain. For more details, visit our [openlm.io](https://openlm.io) or check out our models on Hugging Face.