GPT-5.6 Luna pricing rules
Official list rate; undocumented cache and special tiers are omitted from calculations.
Official sourceCurated cost comparison
GPT-5.6 Luna vs Claude Haiku 4.5 API cost
| Fact | GPT-5.6 Luna | Claude Haiku 4.5 |
|---|---|---|
| Input / MTok | $1.00 | $1.00 |
| Output / MTok | $6.00 | $5.00 |
| Cached input | Not recorded | $0.1000 |
| Context | 1,050,000 | 200,000 |
| Example request | $0.0220 | $0.0200 |
| Verified | 2026-07-16 | 2026-07-16 |
Claude Haiku 4.5 pricing rules
Fastest current Claude tier; cache reads and writes use distinct rates.
Official sourceFrequently asked questions
Which costs less for the example workload?
Claude Haiku 4.5 has the lower list-price result for 10,000 input and 2,000 output tokens. This is not a quality ranking.
Does context size change the result?
It can. Context limits determine eligibility, and documented long-context thresholds can apply higher rates.
Are cached tokens included?
The baseline example excludes cache so the standard rates are comparable. Use the calculator to enter an expected cache hit.
Does batch pricing apply?
Only when the provider documents batch eligibility and the workload can tolerate asynchronous processing.
Are tools included?
No. Search, code execution and other server-side tools can carry separate provider charges.
Which model is best?
Price alone cannot determine best. Evaluate capability, latency, reliability, data policy, availability and task-specific performance separately.