We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
Hacker News·27d ago·5 min read
We made Grok 4.5, GPT-5.5, and Claude build the same apps
TL;DR
Grok 4.5, GPT-5.5, and Claude Opus 4.8 were tasked with building interactive apps, including a 3D Rubik's Cube. Each model produced a unique solution, with Grok 4.5 emerging as the most effective.
✦ Why It Matters
Engineers should evaluate Grok 4.5 for projects requiring efficient and effective interactive app development.
Key Takeaways
- Test Grok 4.5 for building interactive applications.
- Compare performance metrics of different AI models.
- Explore the capabilities of AI in real-world coding tasks.
- Consider cost-effectiveness when selecting AI tools.
Full Summary
Grok 4.5 was recently launched by xAI, touted as their most advanced model for coding tasks. In a competitive build-off, Grok 4.5, GPT-5.5, Claude Opus 4.8, and Claude Fable 5 were given identical prompts to create interactive apps without external libraries or network calls.
The first challenge involved building a colorful, animated 3D Rubik's Cube with scramble and solve functionalities. Each model's output was tested in a real browser environment, with Grok 4.5 demonstrating superior performance in both functionality and cost-effectiveness.
Notably, Grok 4.5 successfully rendered the app with minimal latency, while other models faced challenges. This exercise provides insights into the capabilities of different AI models in practical coding scenarios, emphasizing the importance of model selection for specific tasks.