We’re introducing new developer controls in the GPT-5 series that give you greater control over model responses—from shaping output length and style to enforcing strict formatting. Below is a quick overview of the latest features:
| # | Feature | Overview | Values / Usage |
|---|---|---|---|
| 1. | Verbosity Parameter | Lets you hint the model to be more or less expansive in its replies. Keep prompts stable and use the parameter instead of re-writing. | • low → terse UX, minimal prose.• medium (default) → balanced detail.• high → verbose, great for audits, teaching, or hand-offs. |
| 2. | Freeform Function Calling | Generate raw text payloads—anything from Python scripts to SQL queries—directly to your custom tool without JSON wrapping. Offers greater flexibility for external runtimes like:• Code sandboxes (Python, C++, Java, …)• SQL databases• Shell environments• Config generators | Use when structured JSON isn’t needed and raw text is more natural for the target tool. |
| 3. | Context-Free Grammar (CFG) | A set of production rules defining valid strings in a language. Each rule rewrites a non-terminal into terminals and/or other non-terminals, independent of surrounding context. Useful for constraining output to match the syntax of programming languages or custom formats in OpenAI tools. | Use as a contract to ensure the model emits only valid strings accepted by the grammar. |
| 4. | Minimal Reasoning | Runs GPT-5 with few or no reasoning tokens to minimize latency and speed time-to-first-token. Ideal for deterministic, lightweight tasks (extraction, formatting, short rewrites, simple classification) where explanations aren’t needed. If not specified, effort defaults to medium. | Set reasoning effort: "minimal". Avoid for multi-step planning or tool-heavy workflows. |
Supported Models:
Supported API Endpoints
Note: We recommend to use Responses API with GPT-5 series of model to get the most performance out of the models.
Let's begin with updating your OpenAI SDK that supports the new params and tools for GPT-5. Make sure you've set OPENAI_API_KEY as an environment variable.
!pip install --quiet --upgrade openai pandas && \\
echo -n "openai " && pip show openai | grep '^Version:' | cut -d' ' -f2 && \\
echo -n "pandas " && pip show pandas | grep '^Version:' | cut -d' ' -f2
openai 1.99.2
pandas 2.3.1
The verbosity parameter lets you hint the model to be more or less expansive in its replies.