List your company

Maze Navigation on Maze (Standard)

0.9961AccuracyGPT-5
-0.0398440.2291030.498050.766997Nov 28, 2025

Evaluation results

MethodLinks
GPT-5category=Proprietary Models2025.110.9961
OpenAI o3category=Proprietary Models2025.110.957
Claude 4.5 Sonnetcategory=Proprietary Models2025.110.9414
OpenAI o4-minicategory=Proprietary Models2025.110.9375
WMAct2025.110.8814
PPO - Interactivemode=interactive2025.110.8374
Gemini 2.5 Procategory=Proprietary Models2025.110.8359
PPO - EntirePlanmode=single-turn output2025.110.7504
Qwen3-14Bcategory=Opensource Models2025.110.7227
Qwen3-8Bcategory=Opensource Models2025.110.6367
GPT-4ocategory=Proprietary Models2025.110.0469
Qwen3-8B-Ownbackbone=Qwen3-8B2025.110.0195
Qwen2.5-32B-Instructcategory=Opensource Models2025.110.0068
Qwen2.5-7B-Instructcategory=Opensource Models2025.110