In a study, AI model OpenScholar synthesizes scientific research and cites sources as accurately as human experts

washington.edu

February 6, 2026

4 min read

🔥🔥🔥🔥🔥

32/100

Summary

OpenScholar synthesizes scientific research and accurately cites sources, achieving a level of accuracy comparable to human experts. In a study, researchers found that the AI model significantly reduces the issue of hallucination seen in other models like GPT-4o, which fabricated 78-90% of its outputs.

Key Takeaways

OpenScholar is an open-source AI model specifically designed to synthesize current scientific research and cite sources accurately, achieving performance comparable to human experts.
In tests, OpenScholar outperformed existing AI models, including OpenAI’s GPT-4o, and was preferred by scientists 51% of the time over human-written responses.
The model was trained on a dataset of 45 million scientific papers and utilizes retrieval-augmented generation to incorporate emerging research into its responses.
The project’s code, data, and demo are publicly available and free to use, addressing the need for a transparent system that aids scientists in keeping up with vast amounts of research.

Read original article

View story →