Claude Opus 5.5 is Here! Is Claude Finally Back? (5 Use Cases Tested)

Summary
Peter Yang reviews and tests Anthropic's Claude Opus 5.5, evaluating how it addresses issues from Claude Opus 5, such as overly judgmental personality and repetitive phrases ("slop"). He demonstrates multiple generative workflows, including 3D world creation via Blender and WebGL, digital painting, computer-use drawing, UI/UX mobile app design, automated video editing, and personality self-reflection comparisons against OpenAI's GPT-6 Astra and older Claude models.

What is shown

• [01:06 - 02:31] 3D Golden Gate Bridge Generation: Inspired by Sharif Shameem's GPT-6 Astra recreation of the Palace of Fine Arts, Yang prompts Claude Code to script a 3D flyover of the Golden Gate Bridge in Blender, rendering a dusk scene with traffic.
• [02:32 - 02:57] Comparison with GPT-6 Astra: Shows GPT-6 Astra's generation of the same Golden Gate Bridge prompt rendered in daytime.
• [02:58 - 05:04] 3D "Skyward" Interactive Disney Ride: Claude builds a browser-based WebGL simulation ("Skyward") inspired by Disney's Soarin' Over the World, procedural flights over the Pennine Alps, Greenland icefjords with northern lights, Giza pyramids, Fiji atolls, the Great Wall of China, and Paris at night with fireworks and music.
• [05:05 - 07:32] Claude Painting and Anime Drawing Apps: Interactive web artifacts where Claude paints an Impressionist piece stroke-by-stroke ("Watch Claude Paint") and draws an anime character step-by-step from a photo prompt.
• [07:33 - 08:39] Computer Use Drawing Test: Testing Claude's live browser control to draw Yang's profile picture using basic geometric shapes in a web Paint canvas, alongside Astra's attempt.
• [08:40 - 11:32] Mobile App UI Design via Claude Code: Using the /design command in Claude Code to iterate on UI wireframes and simplify the user onboarding flow for Yang's fitness app (Stronger).
• [11:33 - 13:47] Video Editing via HyperFrames: Utilizing Claude with HeyGen's open-source HyperFrames framework to automatically script video effects, captions, image overlays, animated GIFs, and sensitive data blurring frame-by-frame.
• [13:48 - 15:29] Personality Self-Reflection Comparison: Side-by-side output evaluation of Claude Opus 5 versus Claude Opus 5.5 when prompted to analyze user chat history and provide candid personal feedback.

Claims & numbers

• Generating the Golden Gate Bridge flyover script took roughly 30 minutes to set up in Claude and another 30 minutes to render [02:01].
• Generating the interactive 3D "Skyward" WebGL ride took approximately one hour in Claude [04:45].
• The presenter claims Claude Opus 5 often became overly judgmental and relied heavily on generic phrases ("claudespeak" or "slop") like "here's the honest truth", whereas Claude Opus 5.5 produces more direct, actionable feedback [00:20, 14:10, 14:57].
• The presenter notes Claude still lacks an integrated image generation model, requiring external tools (like ChatGPT) to produce raster assets for app mockups [11:12].

Notable quotes

• "Well, I'm happy to share that the latest Claude Opus model fixes a lot of these problems. And some of what it can do is just amazing to see." — Peter Yang [00:30]
• "I think the TL;DR here is that both Claude and GPT have essentially solved 3D model generation." — Peter Yang [02:47]
• "For the first time in a long time, I think Claude feels like Claude again." — Peter Yang [16:09]

Assessment
This is an independent user review and hands-on capability demonstration of Claude Opus 5.5 across complex coding, 3D scripting, UI design, and agentic tasks. While the presented outcomes are genuine working projects and artifacts, generation and rendering times are expedited through video cuts and time-skips rather than demonstrated entirely in real time.

Described by gemini-3.8-flash on 2026-09-29 from the video's audio and frames.