fix: report prompt_tokens correctly for LLM models in SimpleEngine#236

waybarrios merged 2 commits intowaybarrios:mainwaybarrios/vllm-mlx:mainfrom

sjswerdloff:fix/prompt-token-countingsjswerdloff/vllm-mlx:fix/prompt-token-countingCopy head branch name to clipboard