Is this better than Llama3.1 or GPT-4o? Mistral just released their Large 2 LLM model