Mirendil has announced the release of its inaugural model, Fern 0.1, which features a remarkable 10M-token context window. This model has been pretrained from scratch and is specifically targeted at advancing research in agentic AI.
Fern 0.1 has achieved a perfect score on the RULER single needle test at the 10M-token context, showcasing its potential for high-level performance. However, it currently lags behind leading models in the area of agentic coding, indicating room for further development and refinement.
The introduction of Fern 0.1 marks a significant step for Mirendil as it seeks to contribute to the growing field of agentic AI, which emphasizes the development of autonomous systems capable of complex decision-making.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
