In my experiments, its non-reasoning variant is the strongest non-reasoning LLM in its size class, and outperforms much larger LLMs like Llama 4 Maverick in programming and mechanical engineering problems.
In my experiments, its non-reasoning variant is the strongest non-reasoning LLM in its size class, and outperforms much larger LLMs like Llama 4 Maverick in programming and mechanical engineering problems.