List your company

Task Success on SciWorld

68.21RealGPT-5
10.417225.421140.42555.4289Dec 21, 2025

Evaluation results

MethodLinks
GPT-5World Model=Qwen2.5-7B WorldModel2025.1268.2164.161.030.89
GPT-5World Model=Llama3.1-8B WorldModel2025.1268.2163.0757.440.84
Claude-sonnet-4.5World Model=Qwen2.5-7B WorldModel2025.126645.6457.950.88
Claude-sonnet-4.5World Model=Llama3.1-8B WorldModel2025.126658.4653.330.81
Gemini-2.5-flashWorld Model=Qwen2.5-7B WorldModel2025.125639.4945.640.82
Gemini-2.5-flashWorld Model=Llama3.1-8B WorldModel2025.125657.9552.310.93
GPT-4.1World Model=Qwen2.5-7B WorldModel2025.1243.4145.7946.321.07
GPT-4.1World Model=Llama3.1-8B WorldModel2025.1243.4145.1335.380.82
GPT-4-turboWorld Model=Qwen2.5-7B WorldModel2025.1236.795036.60.99
GPT-4-turboWorld Model=Llama3.1-8B WorldModel2025.1236.7952.3144.11.2
GPT-4oWorld Model=Qwen2.5-7B WorldModel2025.1234.9737.6331.440.9
GPT-4oWorld Model=Llama3.1-8B WorldModel2025.1234.9737.6332.990.94
GPT-4o-miniWorld Model=Qwen2.5-7B WorldModel2025.1212.6412.048.90.7
GPT-4o-miniWorld Model=Llama3.1-8B WorldModel2025.1212.6410.788.210.65