Inherent, founded by DeepMind alumni, says its AI ‘teammate’ just outperformed Anthropic and OpenAI at replicating research
Built by DeepMind alumni, British AI lab Inherent released Faraday, an AI agent whose ability to replicate scientific papers could be a stepping stone for innovation.
Inherent, a London-based AI lab founded by former DeepMind employees, has unveiled its AI assistant, Faraday, which has outperformed larger models from Anthropic and OpenAI in independently reproducing findings from scientific papers. Despite being a relatively lesser-known startup compared to its better-funded rivals, Inherent has attracted $50 million in seed funding and showcased the capabilities of Faraday, an AI agent designed to replicate research findings.
Faraday's success was driven by its unique training method, reinforcement learning, which enabled the AI to develop "research taste" – an ability to decide which experiments are worth conducting and how to design them effectively. Measuring against Anthropic's Claude Opus 4.8 and OpenAI's GPT-5.5, both larger models, Faraday used a smaller model, Qwen 3.6, with only 27 billion parameters, demonstrating its efficiency.
Rather than focusing on coding tools, Inherent aimed to create an AI scientist that can contribute across multiple scientific fields, steering clear of building agents that merely provide users with desired responses. The team, composed of 12 employees working together in an office in King's Cross, aims to expand its workforce to 20-25 people by the end of the year, making Inherent a promising destination for DeepMind alumni considering a change of scenery.
Written by urgent.news from TechCrunch's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
