Build next-gen apps with gpt-5.4-mini
Pricing
TokenLab Price
$0.225
Discount: 70%
| Official Price | TokenLab Price | Discount | |
|---|---|---|---|
| Input | $0.75 | $0.225 | 70% |
| Output | $4.50 | $1.35 | 70% |
| Cache Read | $0.075 | $0.0225 | 70% |
One-click test
Sign in once and Web Agent keeps this model, prompt, and request preset for you.
Test gpt-5.4-mini in Web Agent with a short request at /v1/responses, then show request body, latency, and response.
API workbench
The default route for production. The code sample below uses this endpoint with the format you pick.
curl https://api.tokenlab.sh/v1/responses \
-H "Content-Type: application/json" \
-H "Authorization: Bearer sk-xxx" \
-d '{
"model": "gpt-5.4-mini",
"input": "Hello!"
}'Use cases
Best forVision
Reading images, parsing documents, and answering visual questions
Agents and tools
Drive reasoning, support triage, tool calls, and multi-step task flows.
Developer workflows
Generate, review, or debug code without rewiring your stack.
Knowledge assistants
Ship chat, search, and retrieval with predictable cost and behavior.
Side-by-side test
See real response quality, latency, and price before this becomes your production default.
Prompt examples
Write a concise support reply and list the assumptions behind it.
Review this API design and call out the top three integration risks.
Turn a long changelog into release notes a non-engineer would read.
Cost Calculator
FAQ
How much does gpt-5.4-mini cost?
On TokenLab, gpt-5.4-mini costs $0.2250 per 1M input tokens and $1.3500 per 1M output tokens. Cache and per-request prices show in the pricing table when they apply.
What is gpt-5.4-mini best for?
gpt-5.4-mini is a strong fit for Vision, Tool Use, Prompt Cache. You can call it through TokenLab with one API key.
How do I call the gpt-5.4-mini API?
Get a TokenLab API key, then send your request to https://api.tokenlab.sh/v1/responses. The API workbench above has a recommended endpoint and copy-ready code.
Which endpoint should gpt-5.4-mini use?
Use https://api.tokenlab.sh/v1/responses as the default for gpt-5.4-mini. If a provider-native format is supported, the API workbench shows that endpoint too.
Can I test gpt-5.4-mini before integrating it?
Yes. Try in Web Agent opens a ready test for gpt-5.4-mini and keeps your prompt after sign-in, so you don’t lose context.
