Claude Fable 5
16 published runs · succeeded in this ranking by Claude Fable 5.1
Benchmark archive
Inspect published runs from superseded models and tests outside the active suite.
Superseded model versions
GPT-5.6 Sol
10 published runs · succeeded in this ranking by GPT-6 Sol
Grok 4.6
16 published runs · succeeded in this ranking by Grok 4.7
Claude Opus 5
11 published runs · succeeded in this ranking by Claude Opus 5.5
GPT-5.6 Terra
9 published runs · succeeded in this ranking by GPT-6 Sol
GPT-5.6 Luna
9 published runs · succeeded in this ranking by GPT-6 Luna
GPT-5.5
9 published runs · succeeded in this ranking by GPT-6 Sol
DeepSeek V4 Flash
11 published runs · succeeded in this ranking by DeepSeek V4.1 Flash
Grok 4.5
7 published runs · succeeded in this ranking by Grok 4.7
Retired visual tests
Jabberwock
22 historical runs
Low Poly World
22 historical runs
Office Life
22 historical runs
Petri Dish
22 historical runs
Universe Simulator
23 historical runs
Vice City
22 historical runs