Playground · Model Advisor
Estimating long-context cost
How large windows and cache tiers affect monthly cost—avoid extrapolating from short-chat rates. Confirm current list prices on vendor sites; estimate with Model Advisor using your monthly token volume.
Models: kimi · claude · gemini · qwen
Open Model Advisor →Public pricing-tier notes
- Longer windows weight input tiers more; repeated prefixes may qualify for cache pricing.
- RAG can be cheaper than pure long context—weigh failure cost too.
- Confirm current list prices on vendor sites; estimate with Model Advisor using your monthly token volume.
For estimates with your monthly token volume, go to Model Advisor. Official vendor prices are authoritative; on-site figures are estimates only.
FAQ
Why not multiply by max window size?
Billing follows actual tokens and tier rules; max window ≠ every request fills it. Estimate with real monthly usage.