The Smallest AI Model Won! (Claude Coding Challenge) - ASCII Art Coding Challenge

Which Claude model writes the best code? Episode 2 of the LLM Coding Challenge — this time with a harder prompt: generate ASCII-art boxes around each character of a string. Same prompt. One shot. No follow-ups. We score each model across 8 criteria: - Correct output format - Handles empty string - Code visible in response - Example output shown - Multiple test cases - Docstring included - Clean code - Explanation provided The result? Haiku wins with a perfect 8/8. Opus scores 6/8, and Sonnet comes in at 5/8. The smallest model delivered the most complete response — with a docstring, empty string handling, and full code visibility. All code was generated using Claude Code CLI with the -p (one-shot) flag. Source code: https://github.com/GoCelesteAI/ascii_art_boxes Next episode: an even harder challenge. Subscribe to see if Haiku can keep the crown. --- Chapters: 0:00 Intro 0:15 The Prompt 0:30 Opus 1:18 Sonnet 2:06 Haiku 3:04 Scorecard 3:39 Outro --- #Claude #AI #CodingChallenge #LLM #Opus #Sonnet #Haiku #Python #Programming #AICode #ASCIIArt Tags: Claude, AI, coding challenge, LLM, Opus, Sonnet, Haiku, Python, ASCII art, programming, AI code, Claude Code, Anthropic, code comparison, AI models