By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.

Unleash Innovation Through Open Collaboration

Welcome to DiscoResearch, where we're revolutionizing AI research through the power of open collaboration. Join us on our journey as we push the boundaries of language modeling and beyond.

AI Models

Our AI Models have been downloaded over 500.000 times

Through our open-source work in DiscoResearch, we contribute on many fronts to advancing the state of LLM development. Take a look at our models.

Llama 3 DiscoLM 8b v0.1 Experimental
Llama 3 DiscoLM German 8b v0.1 Experimental is an experimental Llama 3 based version of DiscoLM German. This is an experimental release.
DiscoLM Mixtral 8x7b alpha
An experimental 8x7b MoE model based on Mistral AI´s Mixtral 8x7b. This model is based on experimental code converting the model weights to huggingface format.
DiscoLM German 7b
DiscoLM German 7b is a Mistral-based large language model with a focus on German-language applications and the successor of the EM German model family.
DiscoLM 120b (Alpha)
DiscoLM 120b (Alpha) is an experimental 120b model based on Alpindale´s Goliath 120b, a merge of different Llama2-70b models.
Mixtral 7b 8 Expert
This is a preliminary HuggingFace implementation of the newly released MoE model by MistralAi. Make sure to load with trust_remote_code=True.

Pioneering Open-Source
Research in Language Models

Innovating and Fine-Tuning Language Model Evaluation Techniques

As an open source research collective, we are on a mission to accelerate research on large language models through open collaboration and free information sharing. We believe that transparency and joint efforts lead to the best results.

Breaking Barriers: Driving
Innovation Through Collaboration

We actively collaborate with other research communities, such as Hessian.AI, LAION, AlignmentLab AI, DFKI, and others to pool resources, share insights, and advance the collective understanding of LLMs.

Pioneering Open Collaboration in modern LLM AI Research

We believe in the power of an inclusive environment and encourage researchers from diverse backgrounds to share ideas and engage in joint projects. By fostering an open exchange, we want to accelerate innovation and drive positive change in the AI community.


Join us on our AI Journey

Expertise in
Pre-Training and Fine-Tuning

We have significant expertise in continuous pre-training and fine-tuning of language models, allowing us to optimize and adapt models for specific tasks and languages.

Focus on non-english models

One of our main areas of focus is the research and development of non-English language models that address the needs of diverse linguistic communities.

advanced evalution techniques

We have significant expertise in continuous pre-training and fine-tuning of language models. By leveraging the latest methodologies, we optimize and adapt models for specific tasks and languages.

open-source datasets

We are constantly working on developing and publishing novel and diverse open-source datasets our community can use to improve existing and new models.


Most Asked Questions

Find answers to frequently asked questions about DiscoResearch.
If you still can’t find your question here, feel free to contact us.

What is DiscoResearch's primary focus?


DiscoResearch is an open-source research collective primarily focused on the research and training of large language models (LLMs). Our focus areas include non-English models, advanced evaluation techniques, and reinforcement learning (RL) techniques.

What sets DiscoResearch apart in the field of LLM research?


We have extensive experience in continued pre-training and fine-tuning, particularly of non-English models. As of the latest count, our models have been downloaded more than 500,000 times, showcasing the trust and recognition we've earned in the community.

How does DiscoResearch collaborate with other communities?


DiscoResearch aims to connect researchers from different communities to work together on common goals in the field of large language models. We are in close collaboration with initiatives like hessian.AI, LAION, AlignmentLab AI, and other similar communities to leverage each other's expertise and resources.

How can individuals contribute or get involved with DiscoResearch?


If you're interested in collaborating with DiscoResearch or have any questions or feedback, we encourage you to reach out to us on our Discord server. Whether you're an experienced researcher or just starting out, we welcome contributions and participation from anyone passionate about advancing the boundaries of open LLMs.

Contact us

Contact the DiscoResearch team

Send us a message with your request or join our 800+ members Discord server.