Large Language Models (LLMs) have demonstrated strong performance in Natural Language Processing (NLP), yet their evaluation in Arabic—particularly low-resource varieties like the Tunisian Dialect (TD)—remains a pressing need. This study investigates data augmentation using GPT-4o, Arabic-pretrained models, and LLMs with few-shot learning for spontaneous Speech Emotion Recognition (SER). We employ both open-source and proprietary Automatic Speech Recognition (ASR) systems to generate transcriptions of TD speech, which are then used as input features for SER models. Our findings show that the best performance was achieved with AraBERT, reaching an accuracy of 77% and outperforming all alternatives. These results highlight the critical role of ASR quality in downstream SER tasks, as well as the importance of selecting appropriate models for underrepresented languages, with implications for human-computer interaction and mental health applications.