GPT-5 mini token cost & pricing

GPT-5 mini is the tier most production chat products should default to. It keeps the 400K context window and the reasoning framework of the flagship at $0.25 per million input tokens and $2 per million output — a fifth of GPT-5 on input and a fifth on output, which makes the blended per-request cost roughly 80% lower.

GPT-5 mini specifications

What GPT-5 mini costs in practice

Four common workloads, priced at list rates. Monthly figures assume 30 days.

How many requests does $100 buy?

Assuming a 1,000-token prompt and a 500-token response per request.

How GPT-5 mini pricing actually works

The pricing structure mirrors GPT-5 exactly, just scaled down: 8x from input to output, cached input at $0.025 per million. Because scale-down is uniform, the routing decision between GPT-5 and GPT-5 mini is purely about quality on your specific task, not about where the money goes. Test on your real data, not on benchmarks.

When to use GPT-5 mini

Use it for customer support chat, summarization, structured extraction, moderation assistance and most RAG question answering. It is the right starting point and often the right ending point — many teams only discover they need the flagship after a measurable quality gap on a specific task class.

How to cut your GPT-5 mini bill

The mini tier already removes most of the savings pressure, so the remaining wins are architectural: cache the static part of your prompt, keep conversation history truncated to what the model actually needs, and batch offline jobs rather than calling per-item.

Other OpenAI models

Cheaper alternatives to GPT-5 mini

• Llama 4 MaverickMeta50% cheaper
• GPT-4o miniOpenAI67% cheaper
• DeepSeek V3.2 (chat)DeepSeek69% cheaper
• GPT-5 nanoOpenAI80% cheaper