OpenAIGPT-5.4 pricing
$2.50 per 1M input tokens and $15.00 per 1M output tokens. In real work that is 1.8¢ per chat answer and 4.0¢ to summarize a 20-page PDF.
What GPT-5.4 costs per task
Token counts estimated from a measured model of the same lab (GPT-6 Luna), rescaled by tokenizer. Reasoning tokens are billed as output.
| Task | Cost | Tokens used | Price rank | Basis |
|---|---|---|---|---|
| Answer a chat question | 1.8¢$17.80 / 1,000 | 45 in · 540 out · 639 reasoning | 299 of 336 | estimated |
| Reply to a customer email | 0.21¢$2.06 / 1,000 | 258 in · 94 out | 234 of 336 | estimated |
| Write a 1,500-word blog post | 3.6¢$35.55 / 1,000 | 59 in · 2,217 out · 143 reasoning | 284 of 336 | estimated |
| Write 10 product descriptions | 1.6¢$16.02 / 1,000 | 283 in · 987 out · 34 reasoning | 251 of 336 | estimated |
| Translate 1,000 words | 2.4¢$24.25 / 1,000 | 1,330 in · 1,328 out · 67 reasoning | 258 of 336 | estimated |
| Proofread one page (500 words) | 1.3¢$13.12 / 1,000 | 718 in · 664 out · 91 reasoning | 284 of 336 | estimated |
| Summarize a 20-page PDF | 4.0¢$40.46 / 1,000 | 12,987 in · 449 out · 84 reasoning | 287 of 336 | estimated |
| Extract data from an invoice | 0.94¢$9.36 / 1,000 | 373 in · 422 out · 140 reasoning | 274 of 336 | estimated |
| Analyze a 200-row spreadsheet | 5.4¢$53.86 / 1,000 | 4,779 in · 456 out · 2,338 reasoning | 244 of 336 | estimated |
| Answer from 5 retrieved documents (RAG) | 1.4¢$14.29 / 1,000 | 3,690 in · 234 out · 104 reasoning | 269 of 336 | estimated |
| Classify 100 support tickets | 1.9¢$18.68 / 1,000 | 1,261 in · 445 out · 590 reasoning | 255 of 336 | estimated |
| Review 500 lines of code | 9.0¢$90.26 / 1,000 | 4,557 in · 337 out · 4,921 reasoning | 285 of 336 | estimated |
| Write a SQL query | 1.2¢$12.04 / 1,000 | 126 in · 266 out · 516 reasoning | 285 of 336 | estimated |
| Run a 50-step coding agent | $1.07$1,073 / 1,000 | 2,237,500 in · 15,000 out · 7,000 reasoning | 240 of 336 | modeled |
Price at every provider
List prices per 1M tokens (input / output) at each host serving GPT-5.4 through OpenRouter today. Hosts can differ in quantization, context and speed.
| Provider | Input / output | Cached input | Context | Uptime |
|---|---|---|---|---|
OpenAIunknown | $1.25 / $7.50 | $0.12 | 1.05M | |
 Azureunknown | $2.50 / $15.00 | $0.25 | 1.05M | 100% |
OpenAIunknown | $2.50 / $15.00 | $0.25 | 1.05M | 100% |
Azureunknown | $2.75 / $16.50 | $0.28 | 1.05M | |
Azureunknown | $2.75 / $16.50 | $0.28 | 1.05M | 100% |
 Amazon Bedrockunknown | $2.75 / $16.50 | $0.28 | 1.05M | |
OpenAIunknown | $5.00 / $30.00 | $0.50 | 1.05M |
Price history
First seen in our price history on Mar 17, 2026 at $2.50 input / $15.00 output. The list price has not changed since.
FAQ
How much does GPT-5.4 cost?
GPT-5.4 costs $2.50 per 1 million input tokens and $15.00 per 1 million output tokens, with cached input at $0.25 (list price on Oct 18, 2026). In practice one chat answer costs about 1.8¢ and summarizing a 20-page PDF about 4.0¢.
Where is GPT-5.4 cheapest?
Among 7 providers we track, the lowest list price today is at OpenAI: $1.25 input and $7.50 output per 1M tokens.
Does GPT-5.4 charge for reasoning tokens?
Yes. Reasoning (thinking) tokens are billed as output tokens. That is why the cost per task can be much higher than the visible answer suggests - our task costs include them.
Is the GPT-5.4 cost per task measured or estimated?
Estimated: GPT-5.4 was not run directly. Tokens come from GPT-6 Luna and are rescaled by the tokenizer of OpenAI. Real usage can differ.