Inherent Labs Introduces Faraday 27B AI Scientist
The model adds scientific intuition to coding agents for paper replication.
Inherent Labs posted its introduction of Faraday, a 27B-parameter model trained via long-horizon RL. The announcement states the system extends coding agents with scientific intuition and outperforms Claude Opus 4.8 and GPT-5.5 on replicating research papers. Multiple researchers including Jakob Foerster, Roberta Raileanu, Ross Taylor, and Nando de Freitas shared or quoted the post on X. A linked arXiv paper discusses training AI scientists for research replication.
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
Reactions from ranked influencers
42 postsSuper cool progress from @tantumscollins @edwardfhughes and the team at @inherent_labs towards optimizing models for that elusive thing we call "research taste"
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
@inherent_labs' first research result is out! Our AI scientist Faraday learns to serve as a layer of scientific intuition on top of the advanced capabilities of coding agents. Check out the paper here .
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
I’m very excited about the Inherent team - this paper gives a glimpse of what they’ve been working on…👇
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
Really excited for our first public AI scientist artifact 🚀 At Inherent, we are training agents with scientific intelligence. Our agent Faraday learns to replicate research. It outperforms leading LLMs and leverages coding agents.
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
This is really interesting work. Inherent got their agent to reproduce masked out figures in popular ML papers by running the same experiments scaled down, with careful controls in place to maintain construct validity. They GRPO a 27B model to act as a controller for frontier…
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
