Claude Haiku 4.5 vs GPT-6 Luna
A cost comparison of two tracked small tiers. Tier roles do not establish equal capability, quality or speed.
| Model / service | Tier | Input / 1M | Output / 1M | Observation |
|---|---|---|---|---|
| Claude Haiku 4.5 Anthropic | Small | $1.00 | $5.00 | 2026-10-05 Verified |
| GPT-6 Luna OpenAI | Small | $0.1 | $0.5 | 2026-10-05 Verified |
Scroll the table sideways to see every column.
These are recorded base standard list rates in USD per million tokens. Promotions, long-context surcharges, Batch, cache write/storage and other service charges can change the bill. Use the calculator for its supported workload rules and check the linked provider terms. A carried observation is the last known rate, not a fresh verification.
Claude Haiku 4.5: context capacity 200K. Official Anthropic pricing source.
GPT-6 Luna: context capacity 1.05M. Official OpenAI pricing source.
What would 100,000 English words cost?
Using TokenScale’s estimate of 4 tokens per 3 English words, 100,000 words is about 133,333 tokens. Each column below bills only that direction at the recorded base rate; it is not a complete conversation bill. Language, tokenization, context limits and provider terms can change the result. These are calculations, not provider quotes.
| Model | Input only | Output only |
|---|---|---|
| Claude Haiku 4.5 | $0.1333 | $0.6667 |
| GPT-6 Luna | $0.0133 | $0.0667 |
Scroll the table sideways to see every column.
All model comparisons · All tracked models · Verification method