Despite the polish, users have reported three edge cases:
While a "patch" release might seem minor on paper, in the world of Large Language Models (LLMs), the difference between 2.4.0 and 2.4.2 can be the difference between a model that is functional and one that is production-ready. The 2.4 branch introduced advanced "Chain-of-Thought" reasoning, but initial implementations suffered from context drift—where the model would lose track of the initial prompt during long interactions. EVA 2.4.2 is the polished result, ironing out these drift issues to deliver a cohesive reasoning engine. eva 2.4.2
Developers reported that the /api/v2/assets endpoint would time out when listing more than 10,000 objects. EVA 2.4.2 implements cursor-based pagination, replacing the flawed offset/limit method. This change drastically improves response times from 45 seconds to under 200 milliseconds. Despite the polish, users have reported three edge