Pricing

Billed per token, in USD.

Compare on tokens, not on price alone. Most Community models count input tokens more generously and add a fixed overhead to every request, so a short prompt costs more than its unit price suggests. See Usage.

Price per 1M tokens, USD
ModelInputOutput
claude-fable-53.6818.38
claude-fable-5-13.6818.38
claude-opus-4-61.587.88
claude-opus-4-71.587.88
claude-opus-4-81.587.88
claude-opus-51.587.88
claude-sonnet-4-60.954.73
claude-sonnet-50.633.15
gpt-5.50.794.73
gpt-5.6-luna0.160.95
gpt-5.6-sol2.6315.75
gpt-5.6-terra0.392.36
gpt-6-astra1.587.88
gpt-image-21.056.30

Video prices

USD per million billable tokens. Domestic and overseas rates are equal. Choose one reference category; the rates are not added together.

ModelOutput sizeNo reference videoWith reference video
doubao-seedance-2-0-mini-260615480p / 720p$3.382353$2.058824
doubao-seedance-2-0-fast-260128480p / 720p$5.441176$3.235294
doubao-seedance-2-0-260128480p / 720p$6.764706$4.117647
doubao-seedance-2-0-2601281080p$7.500000$4.558824
doubao-seedance-2-0-2601284k$3.823529$2.352941
doubao-seedance-2-5-260628480p / 720p$10.294118$6.176471
doubao-seedance-2-5-2606281080p$11.323529$6.764706

Mini / Fast 1080p pricing is not yet available.

What counts

Input and output tokens are reported in the usage block of every response. Billing is quantised, so a very small request may round to zero.

A request that fails is not billed. A request you disconnect from is — see Errors.