Claude Opus 5.5 Beats ChatGPT 6 Astra in 8 of 12 Real-World Trials

Anthropic's Claude Opus 5.5 outperformed ChatGPT 6 Astra in 8 of 12 practical trials, according to a new comparison. The model handles simulations, game development, writing, design and 3D animation, including a 60-second time-loop game and a 22-second book-to-city animation.

• Claude Opus 5.5 won 8 of 12 real-world trials against ChatGPT 6 Astra
• Created a firefly simulation with adjustable movement and light parameters
• Built a 60-second time-loop game and a 22-second 3D animation
• Main limitation is tasks exceeding its token constraints