Is it nerfed? We regularly give every model the same coding tasks, track whether its work gets worse, and let you know when it does.
Build a water simulation sandbox with its own physics engine, playable on desktop and phone.
Desktop
Agent session
Fable 5.1Claude Code 2.1.28542 minhigheffort7.9Min170Kout$13.45API cost
GPT-6 AstraCodex CLI 0.159.031 minhigheffort2Min43Kout$5.15API cost
Opus 5.5Claude Code 2.1.2851 h 23 minhigheffort64Min430Kout$24.19API cost
Sonnet 5.5Claude Code 2.1.28549 minhigheffort29Min310Kout$9.73API cost
GPT-6.1 SolCodex CLI 0.159.035 minhigheffort1.7Min57Kout$0.90API cost
Grok 4.7Grok Build 1.0.411 h 30 minhigheffort
Sonnet 5Claude Code 2.1.2851 h 30 minhigheffort61Min320Kout$16.37API cost
GPT-6 SolCodex CLI 0.159.033 minhigheffort2.9Min55Kout$1.30API cost
Grok 4.6Grok Build 1.0.461 h 8 minhigheffort18Min230Kout$11.24API cost
Haiku 4.5Claude Code 2.1.2856 minhigheffort2.9Min49Kout$0.62API cost
GPT-6 LunaCodex CLI 0.159.07 minhigheffort420Kin22Kout$0.02API cost
haiku-5.5Claude Code 2.1.29352 minhigheffort35Min400Kout
gemini-3.8-flashAntigravity CLI 1.2.1412 minhigheffort7.9Min150Kout$1.47API cost
gemini-3.1-proAntigravity CLI 1.2.146 minhigheffort190Kin44Kout$0.68API cost
gemini-3.6-flashAntigravity CLI 1.2.147 minhigheffort1.1Min49Kout$0.48API cost
No desktop video