thinking machines' inkling is now visible on openrouter's model board, marking a significant step for the lab. this open-weight multimodal mixture-of-experts model boasts 41 billion active parameters from a total of 975 billion. it's designed for general-purpose reasoning, coding, and agentic tool-use systems, which could have broad implications for various applications in AI.
the stakes are high as inkling enters a competitive landscape filled with advanced models. its unique architecture could give it an edge in tasks requiring complex reasoning and interaction with tools. developers and researchers will be watching closely to see how inkling performs in real-world scenarios and its potential impact on existing solutions.
however, some experts caution that the sheer number of parameters doesn't always translate to better performance, and the real test will be how well inkling adapts to specific use cases and integrates into existing workflows. @OpenRouterAI
