gpt-5.4-mini
$442.50 / month- Input cost
- $262.50
- Output cost
- $180.00
- Per 1,000 calls
- $0.885
- Pricing profile
- standard / short context
Source-tracked model comparison
GPT-5.4 Mini vs GPT-5.4 Nano: how much does the routing tier change API cost?
Compare verified pricing and model limits for choosing a GPT-5.4 routing tier for high-volume requests. The cost example uses a classification and routing workload and does not assume equal model quality.
Direct cost answer: gpt-5.4-mini is estimated at $442.50 per month and gpt-5.4-nano at $120.00 for 700 input tokens, 80 output tokens, and 500,000 monthly requests. gpt-5.4-nano is $322.50 lower under these assumptions. This does not identify a quality winner.
Tracked facts
Pricing and model limits
Prices are USD per 1M tokens under each model's verified default profile.
| Field | gpt-5.4-mini | gpt-5.4-nano |
|---|---|---|
| Provider | OpenAI | OpenAI |
| API model ID | gpt-5.4-mini | gpt-5.4-nano |
| Input / 1M | $0.75 | $0.20 |
| Cached input / 1M | $0.075 | $0.02 |
| Output / 1M | $4.50 | $1.25 |
| Context window | 400,000 tokens | 400,000 tokens |
| Maximum output | 128,000 tokens | 128,000 tokens |
| Accepted input | text, image | text, image |
Example workload
Classification and routing cost scenario
700 input and 80 output tokens per request, 500,000 monthly requests, and 0% cached input.
gpt-5.4-nano
$120.00 / month- Input cost
- $70.00
- Output cost
- $50.00
- Per 1,000 calls
- $0.24
- Pricing profile
- standard / short context
Cost result: gpt-5.4-nano is $322.50 lower per month for these assumptions. This is a price comparison, not a model-quality ranking.
Questions to answer before choosing
- Which requests need the Mini tier rather than Nano?
- Can a confidence threshold escalate only uncertain results?
- How do output length and retries change the tier savings?
What this estimate leaves out
Quality and latency are not assumed equal across models; validate labels on a representative evaluation set.
Latency, reliability, output quality, retries, regional processing, and provider-specific tool charges can change the practical decision.