GPT-5 mini API quick start with a real workflow

What this GPT-5 mini guide covers

If you want a cheaper GPT-5 model without dropping back to an older family, GPT-5 mini is the obvious place to start.

The current GPT-5 mini model page describes it as a faster, more cost-efficient version of GPT-5 for well-defined tasks and precise prompts. The broader models guide also points developers toward GPT-5 mini when lower latency and lower cost matter more than frontier-level reasoning.

This guide is about the original GPT-5 mini API snapshot, gpt-5-mini-2025-08-07.

If you are starting a new project and mostly want the latest flagship quality, start with GPT-5.4 from the current models guide. Reach for GPT-5 mini when lower cost and lower latency are the real reason you are choosing it.

By the end, you will have:

• a first successful GPT-5 mini Responses API call
• a structured support-triage workflow that returns strict JSON

If you want the bigger-picture model first, read my GPT-5.0 API guide. If you want the absolute cheapest GPT-5 variant after this one, open GPT-5 nano next.

Get your API key ready

You need an OpenAI account, a funded API project, and an API key from the API keys page. Keys are shown once, so save yours right away.

Then export it in your terminal.

macOS and Linux:

export OPENAI_API_KEY="sk-..."

Windows Command Prompt:

setx OPENAI_API_KEY "sk-..."

If you use setx, open a new terminal before testing the key.

Send your first GPT-5 mini request

GPT-5 mini uses the same Responses API shape as full GPT-5, but it does not support reasoning.effort: "none". In my live test, the API only accepted minimal, low, medium, or high.

This first request worked for me:

curl -s https://api.openai.com/v1/responses \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5-mini-2025-08-07",
    "input": [
      {
        "role": "user",
        "content": [
          { "type": "input_text", "text": "Say hello in one short sentence." }
        ]
      }
    ],
    "reasoning": {
      "effort": "minimal"
    },
    "text": {
      "verbosity": "low"
    },
    "max_output_tokens": 80
  }'

That exact request completed successfully for me and returned Hello!.

Three details matter:

• the model is pinned to gpt-5-mini-2025-08-07
• reasoning.effort is set to minimal
• text.verbosity stays under text

Build something useful: support triage

GPT-5 mini is a good fit for smaller workflows where the prompt is precise and the output shape matters more than deep open-ended reasoning.

Imagine your app receives this message:

Hi, I was billed twice for my Pro plan today. Please refund the extra charge.

You want GPT-5 mini to:

1. classify the issue
2. set a priority
3. decide whether a human should step in
4. draft a safe reply

That is the kind of workflow where mini makes sense. It is clear, bounded, and easy to validate.

Return strict JSON with a schema

curl -s https://api.openai.com/v1/responses \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5-mini-2025-08-07",
    "instructions": "You triage support messages for a SaaS app. Be cautious. Do not promise actions the billing team has not confirmed. Keep reply_draft to 2 short sentences.",
    "input": [
      {
        "role": "user",
        "content": [
          {
            "type": "input_text",
            "text": "Hi, I was billed twice for my Pro plan today. Please refund the extra charge."
          }
        ]
      }
    ],
    "reasoning": {
      "effort": "minimal"
    },
    "text": {
      "verbosity": "low",
      "format": {
        "type": "json_schema",
        "name": "support_triage",
        "schema": {
          "type": "object",
          "properties": {
            "category": {
              "type": "string",
              "enum": ["billing", "bug", "account", "feature_request", "other"]
            },
            "priority": {
              "type": "string",
              "enum": ["low", "medium", "high"]
            },
            "needs_human": {
              "type": "boolean"
            },
            "reply_draft": {
              "type": "string"
            }
          },
          "required": ["category", "priority", "needs_human", "reply_draft"],
          "additionalProperties": false
        },
        "strict": true
      }
    },
    "max_output_tokens": 220
  }'

That exact request completed successfully for me and returned JSON in this shape:

{
  "category": "billing",
  "priority": "high",
  "needs_human": true,
  "reply_draft": "Sorry about the double charge - I have flagged this to our billing team to investigate. Please reply with the transaction IDs or a screenshot of the two charges so we can resolve this faster."
}

That is immediately useful in an app:

• category can route the ticket
• priority can change queue order
• needs_human can trigger escalation
• reply_draft can prefill the first response

If you want to move this pattern into PHP next, my guide on using OpenAI’s API in PHP with openai-php/client picks up right there.

What changes with GPT-5 mini

The big change is not a new shiny parameter. It is the trade-off.

GPT-5 mini keeps the GPT-5-style API surface, but it is tuned for:

• lower latency
• lower cost
• well-defined tasks
• precise prompts

The current pricing page lists GPT-5 mini at $0.25 input and $2.00 output per 1M tokens, which is much cheaper than full GPT-5. The current model page lists a 400,000-token context window, which gives mini a lot of room even though it is the cheaper tier.

So if your task is structured and repeatable, mini often makes more sense than the full model.

How I would choose reasoning effort on GPT-5 mini

For GPT-5 mini, I would use:

• minimal for extraction, classification, and routing
• low when the task needs a bit more judgment
• medium only when mini is struggling and you still want to stay on this cheaper model
• high sparingly

The key point is that none is not available here, unlike on later GPT-5.1+ frontier models.

When GPT-5 mini is a better choice than full GPT-5

Pick GPT-5 mini when:

• the prompt is precise
• the output schema is simple
• you need lower cost
• you expect higher request volume

Do not assume mini is always the right default, though. If the task is more ambiguous, coding-heavy, or tradeoff-heavy, full GPT-5 or a newer model like GPT-5.4 may still pay for itself in answer quality.

Common mistakes with GPT-5 mini

1. Assuming it supports reasoning.effort: "none"

It does not. In my live test, the API rejected none and only accepted minimal, low, medium, or high.

2. Using mini for vague prompts

Mini works best when the task is tightly specified. If the prompt is fuzzy, the lower cost will not save you from lower-quality outcomes.

3. Forgetting that Nano exists

If your workload is even simpler than this one, such as fast classification or summarization at very high volume, GPT-5 nano may be the better fit.

If GPT-5 mini looks close to what you need, these are the next reads I would keep open:

• See when the full GPT-5 model is still worth the extra cost
• Compare GPT-5 mini with the even cheaper GPT-5 nano
• Move this same structured-output workflow into PHP
• Build a better mental model for what GPT-style models are actually doing