To create a fine-tuning job, the 'SNOWFLAKE.CORTEX.FINETUNE('CREATE', ...y function requires the training data query to return columns named 'prompt' and 'completion'. The 'mistral-7b& model is listed as a supported base model for fine-tuning. Models like 'llama4- maverick' , 'openai-gpt-4.1', 'snowflake-arctic' , and 'claude-4-opuS are not listed as base models available for fine-tuning using this function.