Self-Consistency Majority Voting for Math and Code Evals

For non-reasoning models working on tasks with verifiable answers (such as math or testable code), sampling multiple completions (n) at a nonzero temperature and taking the majority answer remains a valuable evaluation and verification technique.

API Config: model="gpt-4.1", n=5, temperature=0.7 Prompt: Let's work this out in a step by step way to be sure we have the right answer. Solve the combinatorial probability problem below and output the final number inside <final_value> tags for majority voting.