A $1 billion Series A
US-based Mirendil has raised a $1 billion Series A round to develop AI models and the agent software built on them. Kleiner Perkins and Clearlake Capital co-led the round. Battery Ventures, Doerr Capital, Gil & Co, Google, M13, Pear VC and S32 also participated.
What it builds
The company has two core products: Fern, an AI model focused on research tasks such as writing code, debugging, tracking training runs and inspecting datasets; and Lor, a platform that lets AI agents write code, run experiments and manage compute resources. Users can hand complex tasks to multiple agents.
Goal: big research with small teams
Mirendil aims to let smaller teams, or even a single researcher, carry out work that previously required a large research group or institution. The company will use the funding to open its research and development processes to more scientists and engineers and to grow its models and agent platform.
Founding
Mirendil was founded in 2026 by Behnam Neyshabur, Harsh Mehta, Shayan Salehian and Tara Rezaei.
