OpenAI

GPT-5 nano pricing

OpenAI's cheapest GPT-5 tier. $0.05 input - classify or enrich a million rows for under a dollar.

Input

$0.05/ 1M tok

Output

$0.40/ 1M tok

Context window
400K
Max output
128K
Cached input
$0.005 / 1M
Verified
2026-04-06

GPT-5 nano is the floor of the GPT-5 family: $0.05 per 1M input and $0.40 per 1M output, 400K context window, 128K max output. Among the 25+ models Calcis tracks, only Gemini 2.5 Flash-Lite ($0.10 / $0.40) is cheaper on input - and nano beats even Flash-Lite on input price.

At these rates you can think of nano as “free for classification” in most practical senses: a 1,000-token prompt with a 500-token reply costs about $0.00025, and 100 million of those calls a month totals around $25,000 - an unusual line item on a spreadsheet where the infrastructure calling the API usually costs more than the API itself.

Calcis counts GPT-5 nano input tokens with o200k_base (tiktoken), the same tokenizer OpenAI uses for billing, so there's no drift between the count you see and the one on your invoice.

Estimate your cost on GPT-5 nano

Paste your prompt into the estimator, pick GPT-5 nano, 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 nano cost per request?
A 1,000-token prompt with a 500-token reply costs about $0.00025 ($0.00005 input + $0.0002 output). At 10 million requests a month, that's around $2,500.
What is GPT-5 nano good for?
High-volume classification, routing, log enrichment, simple chat, and any task where you need a capable model but not a careful one. For reasoning, code, or nuanced generation, step up to mini or full.
Is GPT-5 nano cheaper than Gemini 2.5 Flash-Lite?
On input, yes - $0.05 vs $0.10 per 1M. On output, it's a tie at $0.40 per 1M. Pick on ecosystem (OpenAI API vs Google AI Studio) and specific capability tests.
What's the cached input discount on GPT-5 nano?
$0.005 per 1M cached tokens - a 90% discount on the standard $0.05 input rate. For high-volume workloads with repeated prefixes, this takes marginal cost close to zero.

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