Enterprise AI software company offering a development platform and ready-made applications for large organisations
Mirendil
New· 4Frontier AI lab building models and agents for AI research; Series A of $1B announced November 12, 2026, and a first model, Fern, with a 10M-token context window.
TrustList summary
Instead of using AI to assist researchers, Mirendil wants AI to do the research itself, and it has organized the whole company around that bet. It calls itself a frontier lab building systems that excel at AI research and development, with models trained for the job and a lab redesigned around them. The name comes from a word the company defines as friend of precious things, the hidden and undiscovered, meaning friend of discovery.
Lor, a lab made of agents
Lor is the product in private beta, with a waitlist. Its agents write code, run experiments and iterate for days on any machine, bringing up GPUs as needed and splitting a job across dozens or hundreds of parallel agents. People stay in charge of what to try and what to trust, and can follow or stop any agent. Lor runs frontier models from other developers as well as open-source ones, and is the home for Mirendil's own. The company says it runs on Lor every day. Example tasks include overnight ablation runs with a written report and a repository-wide security audit that opens a pull request for each fix.
Fern and the case for long context
Fern is Mirendil's first model, pretrained from scratch and post-trained from a base model called Mirendil-Base. Version 0.1 supports up to 10 million tokens of context, which the company equates to roughly 23 MB of code or half a million lines. Larger base models in the family, at 0.5 to 2 trillion and 2 to 10 trillion parameters, are listed as still in pretraining. Early access is by request.
One researcher, under $10M
The company reports that a single researcher working with agents pretrained its base coding model in 3.5 months using less than $10 million of compute. Its stated goal is to make frontier AI R&D reachable for scientists in fields such as drug discovery, chemistry, biology and robotics, which otherwise must become AI labs themselves.
Key facts
People: described as a small team; a single researcher built the first base model Funding: $1B Series A led by Kleiner Perkins and Clearlake Capital, announced November 12, 2026 Website: https://mirendil.com
Summary written by TrustList from the business's own published information, Nov 2026.
Categories & features
Trust Score
An earned signal from verification, reviews, awards, transparency and engagement - the vendor can't buy it.
Updated 11/13/2026
Request a demo or quote from Mirendil
Reviews
Questions & answers
No questions yet. Be the first to ask about Mirendil.
Mirendil alternatives
Similar listings buyers compare against this one.
Göteborg AI consultancy delivering machine learning, data science and web crawling services to companies
Open-source conversational AI framework and agent platform for teams building customer-facing assistants

Voice AI agents that answer and make calls for contact centres, banks and insurers

Text analytics software that mines customer feedback for meaning, sentiment and intent
AI software that helps power utilities inspect and assess grid assets from images and data
Aidentify is an AI studio building enterprise AI agents, a smartphone assistant and AI strategy services

Enterprise AI company applying data science to planning in manufacturing and complex supply chains
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- More United States of AmericaBack to the ranked category
- Write a reviewHelp other buyers decide
- ProductsSoftware and SaaS with pricing, features, integrations and alternatives.
- ArticlesGuides, comparisons and research from the editorial desk and the community.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- AI ModelsBenchmark scores and community ratings for every major model.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.


