Skip to main content
Expand a provider to view its model prices. Categories follow the API Reference, with BytePlus first. BytePlus Video Enhancement has separate selectors for Standard / Professional, output resolution, and frame-rate tier, showing one per-minute rate for the selected combination. The default is Standard, 1080P, up to 30 fps. Requests support 1080P, 2K, and 4K; additional metered output tiers are listed for billing reference. Final charges use actual output duration and specifications. See the Video Enhancement reference. Token-billed models show their token rates alongside an approximate media price only where a supported specification is available. Estimates use current customer prices, including applicable discounts. Hover over or focus an estimate for its assumptions. Gemini image estimates use 1,120 output tokens for a 1K image. GPT Image estimates use 1024×1024 at medium quality: 1,756 output tokens for GPT Image 2 and 439 for GPT Image 2.5 Flare and Sunburst, based on the OpenAI output-token calculator. Image estimates cover output only; input and text/reasoning charges may apply separately. Other sizes, quality settings, and auto can differ. Video estimates use the selected resolution, an approximately 16:9 aspect ratio, and 24 fps. Video estimates with video input are per billable second, including both input and output duration. A dash (—) means no unit estimate is provided. LLM requests have variable input and output usage, so no fixed per-request price is shown. Models without a verified reference specification also show no unit estimate. Token rates and actual per-image or per-second rates remain available. Estimates do not change billing. Units per $1 uses the unrounded unit price and the same assumptions as the estimate. For rows without a media estimate, it uses the listed billing unit. Each row covers only the charge shown, not any additional input or output charges. BytePlus Video Enhancement shows seconds per $1, converted from its per-minute rate. Cache storage depends on both token volume and storage time, so no single quantity is shown.

Billing behavior

  • Creating a generation task can reserve wallet balance while the task is running.
  • Successful billable requests are recorded once using a stable request or task identifier.
  • Failed or cancelled tasks are released or remain non-billable according to the endpoint lifecycle.
  • Your Dashboard wallet and ledger show the final customer charge.
For Seedance 2.0-series and 2.5, reference audio has no separate audio-input fee and does not add billable input-video seconds or select the video-input rate. The generateAudio output setting does not change the unit price for these models. Final charges still depend on reported video token usage and the applicable model and resolution price, including source/reference video inputs when present. See Seedance reference audio for examples and limits.
The list shows final customer prices only. Model pages describe capabilities, not a fixed price snapshot.
For endpoint-specific billing behavior, review the relevant API reference before sending a production request.