idler, a San Francisco-based frontier data research lab and AI evaluation platform, has raised $9 million in seed funding to advance its work on evaluating and benchmarking frontier artificial intelligence models.
The round was led by Paradigm, with participation from Y Combinator and Long Journey, along with several angel investors, including Manish Chandra, Catheryn Li, Dan Posch, Janine Leger, Lynett Capital, Feross Aboukhadijeh, Nur Bazylbekov, and Smaiyl Makyshov.
idler plans to use the new capital to expand its core machine learning research team, scale its computing infrastructure and accelerate the development of advanced reinforcement learning environments and evaluation datasets designed for frontier AI models.
The startup develops evaluation and benchmarking systems for global AI labs. Its platform combines specialized evaluation suites, programmatic datasets and interactive reinforcement learning environments to test the capabilities and reliability of advanced AI systems.
Rather than focusing only on basic conversational prompts, idler develops testing environments for complex, long-horizon tasks, including:
These environments are designed to provide more realistic and challenging tests for increasingly capable AI models.
Alongside the funding announcement, idler launched ShelfLife, a public benchmark designed to evaluate autonomous financial and strategy agents.
ShelfLife uses a digital twin that replicates the operational data of a live, multi-brand retailer spanning nine physical stores and three e-commerce platforms.
The benchmark is designed to stress-test AI agents on complex retail operations, financial decisions, and strategic planning in a realistic business environment.
With the new funding, idler aims to strengthen its research capabilities, expand computing resources and build more sophisticated evaluation environments for the next generation of frontier AI models.
Revier Therapeutics Launches With €6 Million Seed to Develop New Cardiometabolic Treatments
