Summary
Peter Yang puts Anthropic’s newly released Claude Opus model through five hands-on tests to see whether it lives up to the hype after months of using ChatGPT as his daily driver. He starts with creative generation experiments, prompting Claude alongside Blender to build a 3D flyover of the Golden Gate Bridge and comparing the output side-by-side with a similar video made using GPT Astra, before pushing further with an ambitious recreation of Disney’s “Soaring Over the World” ride complete with generated music and multiple international scenes.
The video also tests Claude’s artistic range, having it draw in impressionist and anime styles, and evaluates its computer-use capabilities by asking it to redraw a profile picture using an online Microsoft Paint tool — comparing the results directly against GPT Astra’s attempt. Beyond the creative demos, Yang shows practical applications for real work, including using Claude’s new /design feature inside Claude Code to generate alternative UI concepts for his personal fitness-tracking app.
Throughout, the video addresses a recurring complaint about earlier Claude Opus versions: an overly judgmental, robotic tone. Yang shares his verdict on whether the new model fixes that personality issue, offering viewers a practical, example-driven look at where Claude currently stands against OpenAI’s competing tools for both creative and productivity workflows.
📺 Source: Peter Yang · Published September 22, 2026
🏷️ Format: Hands On Build
