Mistral’s ‘Magistral’ Reasoning Model Tops GPT-5 on Maths but Stumbles on French Law
Mistral's new open-weight reasoning model beats GPT-5 on competition maths yet underperforms on the European legal tasks it was pitched to handle.
Mistral's new open-weight reasoning model beats GPT-5 on competition maths yet underperforms on the European legal tasks it was pitched to handle.
Have a question, tip, or story idea? We read every message.