Model Risk

Assess model behaviour, risks, and use cases

70

ChatGPT-5o

OpenAI

High Risk
Hallucination68
Bias75
Toxicity65
Privacy70
Approved Use Cases
Customer SupportKnowledge Work SupportCode AssistanceDocument Analysis
Model Information
ProviderOpenAI
Version1.0
Type
Parameters1.8 trillion
StatusProduction
Strengths & Weaknesses

Strengths

  • Excellent reasoning and instruction following
  • Large ecosystem, high compatibility with enterprise workflows
  • Strong safety research commitment

Weaknesses

  • Output drift across releases may cause inconsistent behaviour
  • High cost compared to open-source alternatives
  • Dependent on OpenAI infrastructure availability