Claude built the only game that made me care about the cats.

Same prompt. One shot. No fixes. Claude Opus 5 xhigh: ~28m · ~170K tokens · ~$4.60 Qwen 3.8 Max: ~35m · ~50K · ~$0.60 GPT-5.6 Sol: ~19m · ~22.5K · ~$0.68 Grok 4.6 xhigh: 23:48 · 113K · $0.68 All four built playable

noclipepeAug 215

Here is Opus 5

AOAug 21

Ox Alpha is definitely NOT GLM 5.5, it’s honestly way worse than GLM 5.3

Just ran the benchmarks: Qwen 3.8 27B choked. Opus 5 still claps everything easily

OmedTheVibeCoderAug 213

1/ This week in enterprise AI, one throughline: capability is not the bottleneck anymore.

2/ Opus 5 scored higher, felt harder to work with day to day. Benchmarks up, trust down. 3/ Auto mode is now Claude Code's default. Wild stat: it caught 89% of harmful actions in testing

Randy J RouseAug 211