gpt-5-nano pricing
gpt-5-nano costs $0.30 per 1M input tokens and $1.20 per 1M output tokens as of 2026-09-12. A workload of 10M input and 2M output tokens per month runs $5.40. The cheapest comparable alternative is mistral-small-3 at $0.80 per 1M output tokens.
gpt-5-nano is priced at $0.30 per 1M input tokens and $1.20 per 1M output tokens, with roughly 120 ms to first token and about 100 tokens/sec output throughput.
Last updated . Model pricing is refreshed twice daily.
Price per 1M tokens
Input $0.30 · Output $1.20 · Output/input ratio 4.0x.
- 1M input + 1M output tokens: $1.50
- 10M input + 2M output tokens/month: $5.40
- 100M input + 20M output tokens/month: $54.00
Cheaper alternatives to gpt-5-nano
Same workload, lower output price. Validate quality before switching.
- claude-haiku-4 — $1.00/1M output (17% cheaper)
- llama-4-behemoth — $1.00/1M output (17% cheaper)
- amazon-nova-lite — $0.80/1M output (33% cheaper)
- mistral-small-3 — $0.80/1M output (33% cheaper)
How much are you wasting on gpt-5-nano?
Price lists are the easy part. Run the free waste scanner on your usage export to see which calls should have run on a cheaper tier, which prompts are oversized, and which retries you paid for twice — with dollars recoverable per finding.
Frequently asked questions
How much does gpt-5-nano cost per million tokens?
$0.30 per 1M input tokens and $1.20 per 1M output tokens, as listed on 2026-09-12.
What does gpt-5-nano cost per month at typical usage?
At 10M input and 2M output tokens per month, gpt-5-nano costs about $5.40. At 100M input and 20M output tokens it costs about $54.00.
Is there a cheaper alternative to gpt-5-nano?
Yes. claude-haiku-4 at $1.00/1M output, llama-4-behemoth at $1.00/1M output, amazon-nova-lite at $0.80/1M output, mistral-small-3 at $0.80/1M output. Validate quality on your own evaluation set before switching.
How fast is gpt-5-nano?
Roughly 120 ms to first token and about 100 output tokens per second.