Skip to content
AI IntelligenceNov 11, 2026Practical Tip
Article

LLM 0.37.2 fixed a security issue where a configured OpenAI API key could accidentally be sent to third-party OpenAI-compatible…

…endpoints, e.g. if the key named by api_key_name was never stored or a responses: true model was run with -o chat_completions 1.

AI-generated: summaries written by AI from the linked sources. How we use AI

AI-generatedSource: Simon Willison
01

Source Brief

LLM 0.37.2 fixed a security issue where a configured OpenAI API key could accidentally be sent to third-party OpenAI-compatible endpoints, e.g. if the key named by api_key_name was never stored or a responses: true model was run with -o chat_completions 1.

02

Practical Tip

When pointing a tool at third-party OpenAI-compatible endpoints, verify your API key is actually stored under the configured api_key_name and be careful with responses: true models run with -o chat_completions 1, to avoid leaking your OpenAI key.