I still need to evaluate it for my own workloads, but if you trust the benchmarks, seems about on par in quality vs GLM 5.3 Flash

Source https://artificialanalysis.ai/models/mistral-large-4?total-c...