Mirendil raises $1B in funding to speed up AI model development
Mirendil Inc., a provider of software that speeds up artificial intelligence development, has raised $1 billion in early-stage funding. The startup stated on Thursday that Kleiner Perkins and Clearlake Capital jointly led the Series A investment. They were joined by more than a half-dozen others including Google LLC. Mirendil was launched last year by two […] The post Mirendil raises $1B in…
Mirendil Inc., an AI software provider, has secured $1 billion in early-stage funding from Kleiner Perkins, Clearlake Capital, and over a dozen other investors, including Google LLC. The startup, founded last year by former DeepMind and Anthropic researchers, aims to accelerate AI model development. CEO Behnam Neyshabur and CTO Harsh Mehta bring extensive experience from their time at Google DeepMind and Anthropic, focusing on pretraining, the initial phase of AI development.
Mirendil's first product, Lor, automates model training by deploying hundreds of AI agents to handle tasks. This allows for automatic error correction during evaluation. Recently, Mirendil developed Mirendil-Base, a coding model that can rival leading open-source LLMs, thanks to its automation technology. The model, Mirendil-Base, was created without any input from colleagues, with AI agents writing over 99% of the code and optimizing the training dataset and architecture.
Fern, a specialized coding model built on Mirendil-Base, can handle 10 million input tokens per prompt, equivalent to 10 times the code of Claude Opus 5.5. The company has also previewed upgraded versions of Mirendil-Base with 500 billion to 10 trillion parameters, demonstrating their commitment to creating increasingly sophisticated LLMs.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
