Haiku Beat Opus?! Claude AI Coding Challenge EP03

Which Claude model writes the best code? Episode 3 of the LLM Coding Challenge — this time we generate ASCII-art pyramids with cycling symbols. Same prompt. One shot. No follow-ups. NEW: This episode shows each model's code using bat for syntax highlighting, so you can compare implementations side by side. We score each model across 8 criteria: - Correct output format - Symbol cycling works - Pythonic approach - Usage example shown - Docstring included - Clean code structure - Main guard included - Explanation provided Results: - Opus: 6/8 — Most concise (13 lines), uses .center(), but no docstring or main guard - Sonnet: 7/8 — Full documentation, but manually calculates spaces instead of using .center() - Haiku: 8/8 — Perfect score! Combines .center() with full docstring and clean structure All code was generated using Claude Code CLI with the -p (one-shot) flag. --- Chapters: 0:00 Intro 0:15 The Prompt 0:30 Opus 1:13 Sonnet 2:06 Haiku 3:06 Scorecard 3:41 Outro --- #Claude #AI #CodingChallenge #LLM #Opus #Sonnet #Haiku #Python #Programming #AICode #ASCIIPyramid Tags: Claude, AI, coding challenge, LLM, Opus, Sonnet, Haiku, Python, ASCII art, pyramid, programming, AI code, Claude Code, Anthropic, code comparison, AI models