OpenAI

GPT-5 mini pricing

GPT-5 at 5x less. Handles most chat, summarisation, and routing indistinguishably from the full model.

Input

$0.25/ 1M tok

Output

$2.00/ 1M tok

Context window
400K
Max output
128K
Cached input
$0.025 / 1M
Verified
2026-04-06

GPT-5 mini is the cost-optimised workhorse of the GPT-5 family: $0.25 per 1M input and $2 per 1M output, 400K context window, same tokenizer as the flagship. Most teams should try mini before defaulting to full GPT-5 - measured head-to-head, the capability gap often doesn't justify the 5x price difference on chat, routing, and summarisation workloads.

A 1,000-token prompt with a 500-token reply costs about $0.0013. At 1 million requests a month that's $1,300 - versus roughly $6,500 on the same shape of traffic against full GPT-5. For a lot of production chat backends that's the difference between profitable and not.

Calcis counts GPT-5 mini input tokens with o200k_base (tiktoken), the same tokenizer OpenAI bills against, so the token count on your screen matches the one on your invoice exactly.

Estimate your cost on GPT-5 mini

Paste your prompt into the estimator, pick GPT-5 mini, and see the exact dollar cost - input tokens counted with the provider's own tokenizer, output tokens predicted by our regression model.

Open cost calculator →

Frequently asked

How much does GPT-5 mini cost per request?
A 1,000-token prompt with a 500-token reply costs about $0.00125 ($0.00025 input + $0.001 output). At 1 million requests a month, that's around $1,250.
Is GPT-5 mini good enough for production?
For chat, summarisation, classification, and routing workloads, yes - benchmarks repeatedly show mini at parity with full GPT-5 on these tasks. For complex reasoning, long chain-of-thought, or code generation, test both and compare on your actual workload.
Does GPT-5 mini support the same context as full GPT-5?
Yes - 400K context window and 128K max output, identical to full GPT-5. The difference is model capacity, not plumbing.
What's the cached input discount on GPT-5 mini?
$0.025 per 1M cached tokens - a 90% discount on the standard $0.25 input rate. Applied automatically when prompt prefixes repeat within 5 minutes.

Pricing verified 2026-04-06 from the provider's rate card.