What API parameters does this quiz cover?

It covers the parameters you set on real OpenAI and Anthropic chat completions: temperature, top_p, max_tokens (and Anthropic's max_tokens requirement), stop sequences, frequency_penalty, presence_penalty, the system role, streaming, seed, and response_format. Each question gives a scenario and asks for the right parameter or value.

Do I need an API key to take the quiz?

No. The quiz runs entirely in your browser with no API calls and no key required. It is a knowledge check, so nothing is sent anywhere and your answers are not stored beyond the current session.

Is the quiz accurate for both OpenAI and Anthropic?

Yes. Questions are written to reflect both APIs and call out where they differ — for example, Anthropic requires max_tokens on every request while OpenAI treats it as optional. Explanations note the provider-specific behaviour where it matters.

What score means I understand API parameters well?

Scoring 70% or higher signals a solid working grasp of the core generation parameters. If you score lower, the per-question explanations point you to exactly which concepts — sampling, length limits, penalties, or formatting — to review before retaking.

Fill in the API Parameter Quiz: OpenAI and Anthropic APIs

Test your AI API parameter knowledge

Setting the right parameters on an LLM API call is the difference between deterministic, well-bounded output and a costly, unpredictable response. This quiz drops you into 20 real scenarios from the OpenAI and Anthropic APIs and asks you to pick the correct parameter and value — the kind of decisions you make every time you wire a model into production.

How it works

Each question gives a concrete goal — “make the output fully reproducible,” “stop generation at a delimiter,” “reduce repeated phrases” — and offers a few parameter choices. Select one and the quiz instantly tells you whether it is right and explains the parameter, including where OpenAI and Anthropic behave differently. Your score is tracked across all 20 questions, and a final summary tells you how solid your grasp is. Nothing is sent to any API; it runs entirely in your browser.

Tips for getting full marks

Remember the core distinctions: temperature and top_p both control randomness but you should generally tune only one; max_tokens caps output length (and is required on every Anthropic request but optional on OpenAI); stop sequences halt generation at a string; frequency_penalty and presence_penalty (OpenAI) discourage repetition and encourage new topics respectively; and seed plus temperature 0 is the closest you get to determinism. Keep those straight and the rest follows.