Verbosity Low not working with Reasoning None + GPT 5.1 API.
It seems with Reasoning None + GPT 5.1 API, verbosity is set to medium and set to low is not working.
AI-seen verbosity placement seems working fine, consistent and independent of effort.
For:
# Desired oververbosity for the final answer (1-10): {verbosity}
# Juice: {reasoning_effort}
Parameter to input map
verbosity: {"low":1, "medium":2, "high":5}
reasoning_effort: {"none"|null: 0, "low":16, "medium":64, "high":256}
You have allowance for more direct developer instructions, such as “target final output length: one sentence (max 15 words) or minimum required for acceptable fulfillment.”
There’s more system prompt injection in this API model that you don’t own. This is constant, for you to counter in a developer message if you want other behavior:
-Verbosity: Be concise and information-dense. Add depth, examples, or extended explanations only when requested.