It beats much larger models like R1 and Gemini 1.5 in math, coding, and agentic tasks, while running with far fewer active weights.
Open-source AI just got way stronger. 🧵
It beats much larger models like R1 and Gemini 1.5 in math, coding, and agentic tasks, while running with far fewer active weights.
Open-source AI just got way stronger. 🧵
Talk details: schedule.cppnow.org/session/202...
Talk details: schedule.cppnow.org/session/202...