COMPARE
Prices are what this site actually bills; performance comes from 30 days of real traffic. Updated automatically.
Real settlement price per million tokens; struck figures are upstream list prices.
| Metric | gpt-5.4 | gpt-5.6-luna |
|---|---|---|
| Input /1M | $0.750 | $1.00 |
| Output /1M | $4.50 | $6.00 |
| Cached /1M | $0.075 | — |
| Cache write /1M | — | $1.25 |
| 1M input + 1M output total | $5.25 | $7.00 |
| Endpoints | gpt-5.4 | gpt-5.6-luna |
|---|---|---|
| openai | ✓ | ✓ |
| openai-response | ✓ | ✓ |
| openai-response-compact | ✓ | ✓ |
| anthropic | ✓ | ✓ |
| gemini | ✓ | ✓ |
| openai-alpha-search | ✓ | ✓ |
On CompassApi, gpt-5.4 and gpt-5.6-luna share the same OpenAI-compatible endpoint and the same API key — switching is a one-line model change, so you can run both and keep whichever fits.
client.chat.completions.create(
model="gpt-5.4", # ← "gpt-5.6-luna"
messages=[{"role": "user", "content": "Hello"}],
)