Mistral AI has released a new model: Mistral-Small-3.2-24B-Instruct-2506

Mistral AI has released a new model, Mistral-Small-3.2-24B-Instruct-2506, which demonstrates notable improvements in text processing capabilities, particularly in instruction following, conversational ability, and tone. However, the overall performance gains are relatively modest—for example, it shows about a 3% improvement on MMLU Pro, while gains on GPQA-Diamond and SimpleQA are less than 0.5% and around 2%, respectively. This model is best suited for fine-tuning in specific domains, and as a dense model, it offers a simpler fine-tuning process compared to Mixture of Experts (MoE) models.