Open Science Initiative

Accelerating discovery through open science

Cohere Labs advances AI through transparent research, open-weights models, and powering a global community of 4,500+ researchers who are pushing boundaries and shaping the future of AI.

Join us

Status:

Open

Next deadline:

Applications are reviewed on a weekly basis

Benefits:

Collaborate with fellow researchers
Exclusive events
Contribute to open science progress

Through open collaboration and a wide range of programs and events, we unlock the potential of a global ecosystems of researchers, engineers, and unlock the potential of a global network of researchers, engineers, and thinkers. United by open science, we break silos, challenge limits, and build AI solutions that are stronger, more ethical, and transformative, proving collaboration is the bedrock of true innovation.

Open Science Community Members

Countries represented among members

Virtual events hosted

Distinct community-led programs

Join our community and shape AI with open science

Large group of people sitting together at a long picnic table in a park on a sunny day, listening to a person standing and speaking. Tall trees and green foliage surround the gathering.

The strength of our open science community lies in its diversity and reach. With members spanning the globe we're creating a dynamic ecosystem where ideas flourish and breakthroughs are born.
Find a specialized subgroup where your interests meet collective action. The Cohere Labs Open Science Community connects researchers focused on specific subjects like multilingual AI, safety, and more. When people with shared goals collaborate, individual curiosity transforms into breakthroughs that push the field forward.
Research is real work, and in our Open Science Community you won't do it alone. Active participants find collaborators who can sharpen your methodology, and help you navigate the path to becoming a recognized AI researcher.

Stories from the community

Model bias
Afri-Aya
Expedition Aya

From Pitch to Publication: A Cross-Continental Collaboration

Abdelrahman Zayed (Canada) asked: Do biased language models have biased thoughts? Through Cohere Labs, a team formed across four countries—Swati Rajwal (US), Shivank Garg (India), Reem Abdel-Salam (Egypt), and Zayed. They tested five language models and discovered that bias in thinking steps showed low correlation with output bias. Their paper was accepted to CoLM 2025, proving that open science enables researchers across continents to transform ambitious questions into peer-validated contributions.

Learn more

Illustration of a computer screen displaying a research paper titled ‘Do Biased Models Have Biased Thoughts?’ with the abstract visible. The screen sits against a blue, patterned background.

Community events

We believe in collective intelligence through inspiration and connection. Events led by Cohere Labs staff and community leaders foster idea exchange, collaboration, and global connections that drive scientific breakthroughs.