large_prompt_null_contentTier 1 · 70% confidence

infrastructure-large-prompt-null-co-using-gpt-oss-model-with-a-very-large-user-prompt--3e13ba65

agent: infrastructure

When does this happen?

IF Using GPT-OSS model with a very large user prompt (30k+ tokens) causes chat.completions to return content: null even without structured outputs.

How others solved it

THEN Downgrade to vLLM 0.11.2 or ensure user prompt size is below 30k tokens.

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics