Skip to main content
API usage is metered in credits, drawn from your BudgetPixel account balance.

You pay only for success

Credits are charged when a job reaches succeeded. You are not charged for:
  • jobs that fail or timeout,
  • the time a job spends queued or processing,
  • requests rejected before generation (e.g. invalid input, content policy).

Per-model pricing

Each model has a published price, returned by GET /v1/models:
  • Images are priced per output image (credits_per_generation). Most requests return one image and cost that price. A request that returns several — see multiple images per request — is charged per image, so a 3-image result costs 3 × credits_per_generation.
  • Video is priced per second (credits_per_unit with unit_type: second), so total cost scales with length_seconds. Some models (e.g. SeeDance 2.0) charge a different per-second rate per resolution — GET /v1/models returns a resolution_pricing map, and the cost is that rate × length_seconds.
  • Music is a flat price per track (credits_per_generation) whatever its length — Music 3.0 / 2.6 = 200, Lyria 3 = 100, Mureka V9 = 60 credits — except the models priced by length:
    • Sonilo Music (sonilo-music) makes instrumental/background music of an exact length and bills per second of the requested duration: credits_per_unit = 4 with a 10-second minimum — 10 s = 40, 30 s = 120, 60 s = 240, 180 s = 720 credits.
    • Sonilo Video Music (sonilo-video-music) composes a soundtrack for an input video and bills the video’s measured length, rounded up to whole seconds, at 15 credits/s with a 10-second minimum — a 30 s clip is 450 credits, a 60 s clip is 900. POST /v1/cost returns the per-second rate for this model (the duration is only known once the video is analyzed at submit time).
    • Eleven Music v2.5 (eleven-music-v2.5) makes songs with vocals, or instrumentals, and bills the requested duration (10–300 s, default 60) at 200 credits per minute (credits_per_unit = 200 with unit_type: minute), prorated per second and rounded up, with a 10-second minimum — 10 s = 34, 30 s = 100, 60 s = 200, 180 s = 600, 300 s = 1,000 credits.
  • Sound effects are priced per second (credits_per_unit with unit_type: second) with a 3-second minimum (min_billable_seconds):
    • sonilo-sfx (text) bills the requested duration at 5 credits/s — 5 × max(3, duration): 3 s = 15, 5 s = 25, 10 s = 50, 30 s = 150, 180 s = 900 credits.
    • sonilo-video-sfx (video) bills the input video’s measured length, rounded up to whole seconds, at 15 credits/s — a 4.2 s clip is 5 s = 75 credits, a 60 s clip is 900. POST /v1/cost returns the per-second rate for this model (the duration is only known once the video is analyzed at submit time).
Some image models have parameter-dependent pricing on top of the per-image rule:
  • flux-2-pro prices by output size: 0.5MP = 15, 1MP = 25 (default), 2MP = 50, 4MP = 80 credits per image.
  • grok-imagine-image-2 prices by quality × size — both parameters move the price together: Omitting either parameter uses the cheapest cell (low / 1K = 50 credits). This model has no high quality tier, and the optional input image is free — editing costs the same as generating. Example: {"size": "2K", "quality": "medium"} costs 100 credits.
  • seedream-5.0-pro prices by output size and input images: 1K = 55 (default) or 2K = 110 credits per image, plus 5 credits per reference_images item after the first (the first input image is free). The input fee applies per generated image — each output is its own generation call carrying its own inputs. Example: a 2K request with 3 reference images and num_images: 1 costs 110 + 2 × 5 = 120 credits.
  • seedream-5.0-flash is a flat 25 credits per image at 1K and 2K, and its reference_images (up to 10) are free — a 2K edit with 10 references still costs 25.
  • seedream-5.5 prices by output size and input images: 1K = 26 or 2K = 46 (default) credits per image, plus 10 credits for every reference_images item (the first one included). The input fee applies per generated image — each output is its own generation carrying its own inputs. Example: a 2K edit with 2 reference images and num_images: 1 costs 46 + 2 × 10 = 66 credits.
  • minimax-h3-image prices by output size and input images: 2K = 34 (default) or 4K = 100 credits per image, plus 12 credits for every reference_images item (the first one included), per generated image. Example: a 2K edit with 2 reference images and num_images: 1 costs 34 + 2 × 12 = 58 credits.
  • flux-3-image prices by output size: 768px = 55, 1K = 65 (default), 1.5K = 90, 2K = 130 and 4K = 790 credits per image, the same for text-to-image and editing. Its reference_images (up to 10) and grounding are free. Example: a 2K edit with 10 references and num_images: 2 costs 2 × 130 = 260 credits.
  • flux-3-image-klein (FLUX 3 Klein) prices by output size: 768px = 11, 1K = 12 (default), 1.5K = 15, 2K = 19, 4K = 48 credits per image, the same for text-to-image and edits; reference images are free.
  • nano-banana-2.1 prices by output size: 1K = 45 (default), 2K = 70 and 4K = 150 credits per image, the same for text-to-image and editing. Its reference_images (up to 9) are free. Search grounding is off by default and adds per image: google_search = 10 credits, image_search = 20 credits. Image search runs a web search too, so the two never add up: with image_search: true the extra is 20 whether or not google_search is set. Example: a 2K request with image_search: true and num_images: 2 costs 2 × (70 + 20) = 180 credits.
  • gpt-image-2.6-flare and gpt-image-2.6-sunburst price by quality alone: low = 70, medium = 120 (default), high = 175, xhigh = 340 and max = 675 credits per image, the same at every resolution (up to 4K), for text-to-image and editing. reference_images (up to 16), complexity and background: "transparent" are free. Example: a 4K request with 3 references, quality: "high" and num_images: 2 costs 2 × 175 = 350 credits.
  • minimax-h3 (video) prices per second of output (160 credits/sec at 2K) plus per second of reference-video input: each reference_videos clip is measured and billed at 160 credits per second of input (rounded up on the combined total). reference_images (up to 5) and reference_audios are free. Example: a 5-second generation guided by a 8.2s reference clip costs 5 × 160 + 9 × 160 = 2240 credits.
  • minimax-h3-max (video) prices per second of output by resolution (480p = 65, 768p = 105, 1080p = 210, 2k = 420 credits/sec — 2k on extension only) plus 170 credits per second of video input: every reference_videos clip, or the source video of an extension (billed on its first 15 seconds), rounded up on the combined total. The first two reference_images are free and each further one adds 50 credits; reference_audios and the audio soundtrack are free. Example: extending a 10.2s clip by 5 seconds at 768p costs 5 × 105 + 11 × 170 = 2395 credits.
  • seedance-2.5 (video) prices per second of output by resolution (480p = 150, 720p = 330, 1080p = 750, 4k = 1250 credits/sec) plus per second of reference-video input (480p = 80, 720p = 160, 1080p = 375, 4k = 750 credits per input second, rounded up on the combined total). reference_images (up to 30) and reference_audios are free. Example: a 5-second 720p generation guided by a 8.2s reference clip costs 5 × 330 + 9 × 160 = 3090 credits. In draft mode a draft (draft: true) is priced as a 480p request and its final (from_draft_job_id) as a 1080p request for the draft’s length and reference-video seconds: a 5-second draft costs 5 × 150 = 750 credits and its final 5 × 750 = 3750.
  • seedance-2.5-flash (video) prices per second of output by resolution (480p = 93, 720p = 200 credits/sec) plus per second of reference-video input (480p = 57, 720p = 123 credits per input second, rounded up on the combined total). reference_images (up to 10) and reference_audios are free. Example: a 5-second 720p generation guided by an 8.2s reference clip costs 5 × 200 + 9 × 123 = 2107 credits. There is no draft mode and no 1080p.
  • seedance-2.0 / seedance-2.0-fast / seedance-2.0-mini (video) bill video input the same way: a video to edit (2–15s) and every reference_videos clip are measured and billed at half the model’s output per-second rate for that resolution on top of the output (rounded up on the combined total). Example: a 5-second 480p seedance-2.0-mini edit of a 4s clip costs 5 × 60 + 4 × 30 = 420 credits. Requests without video input are output seconds only; reference images and audio are free.
  • wan-2.7-video prices per second of output by resolution (720p = 100, 1080p = 150 credits/sec). When a video input is supplied the request becomes a video edit, which bills the input clip’s measured duration at the same full per-second rate on top of the output — there is no half-rate discount here. Video edit also takes its output length from the input clip (length_seconds is ignored), so a clip is billed twice over: once as input, once as output. Example: editing an 8.2s clip at 720p costs 9 × 100 + 9 × 100 = 1800 credits. reference_images (up to 4, video edit only) are free. Text-to-video and image-to-video requests are unaffected — output seconds only.
  • wan-3.0-video prices per second by resolution (480p = 60, 720p = 120, 1080p = 240 credits/sec; the high-speed wan-3.0-video-prime tier is 85 / 170 / 340) — for the output and for every second of reference-video input, at the same rate (input rounded up on the combined total). Reference images and reference audio are free; input video duration + length_seconds can’t exceed 30 seconds. Example: a 5-second 720p edit guided by an 8.2s reference clip costs (5 + 9) × 120 = 1680 credits.
  • flux-3-video prices per second of output by resolution (720p = 200, 1080p = 325, 1440p = 380, 4K = 555 credits/sec); audio on or off costs the same. When a video input is supplied the request extends that clip, and only the new seconds are billed, at the extend rate (720p = 285, 1080p = 365, 1440p = 420, 4K = 610 credits/sec); the source clip is free. Example: extending a clip by 5 seconds at 1080p costs 5 × 365 = 1825 credits, and the result is the source followed by the 5 new seconds.
  • kling-4.0 prices per second of output by resolution (720p = 207, 1080p = 310 credits/sec). With any input video (a video to edit, or reference_videos) every billed second costs 124 (720p) or 186 (1080p) credits, and the billed seconds are the output plus the longer of the input videos and the output (input rounded up on the combined total; an edit’s output keeps its clip’s length). Reference images are free. Examples: editing a 5-second clip at 720p costs (5 + 5) × 124 = 1240 credits; a 10-second 1080p video guided by a 4-second reference clip costs (10 + 10) × 186 = 3720.
  • Video upscaling, Topaz (POST /v1/video-upscales/topazlabs-video-upscale) is priced per second of the input video (measured server-side, rounded up; input up to 20s) by target tier: 1080p = 15 (30fps) / 30 (60fps), 4K = 60 / 120 credits per second.
  • Video upscaling, FLUX 3 (POST /v1/video-upscales/flux-3-video-upscale) is priced per megapixel-second of the output: the output frame’s megapixels (1 MP = 1,048,576 pixels) times the clip’s seconds (measured server-side, rounded up to 0.1s; up to 20s), at 95 credits in precise mode and 135 in creative mode. The output frame is the input times upscale_factor (1.5, 2, 2.5 or 3), capped at about 14 MP. One second of 1080p output costs 188 / 267 credits, of 4K 752 / 1068. Example: a 5-second 720p clip at upscale_factor: 3 (3840×2160) in precise mode costs ceil(7.91 × 5 × 95) = 3,758 credits. The price is held while the job runs and charged only on success.
  • Upscalers (POST /v1/upscales/{model}) are priced per image: p-image-upscale by its target resolution tier — 4 MP = 10, 8 MP = 15, 16 MP = 30, 32 MP = 60 credits (input media never adds to the charge) — and clarity-upscaler at a flat 30 credits.
  • Depth maps (POST /v1/depth-maps/video-depth-anything) are priced per second of the input video (measured server-side, rounded up to whole seconds, at least 5; input up to 30s) by size: small = 20, base = 30, large = 40 credits per second. The output resolution, colormap, keep_audio and side_by_side never change the price, and GET /v1/models returns the rates as size_pricing. Example: a 12.4-second clip at large bills as 13 seconds, 13 × 40 = 520 credits; a 3-second clip at small bills as 5 seconds, 5 × 20 = 100.
POST /v1/cost resolves all of this — send the exact body you plan to generate with.
Your plan’s perks follow you to the API. Model discounts and active promotions apply to API charges exactly as they do on the web (POST /v1/cost quotes your discounted price), and the Pro/Ultra daily free-models allowance — a shared pool of free generations per day (Pro 100, Ultra 300) across qualified image models like flux-2-klein and qwen-image — covers API requests too. Once the day’s pool is used, those models charge their standard price until the reset (midnight UTC). Other promotional free daily generations remain web/app-only.

Utility pricing (conversions & posts)

These endpoints aren’t models, so they don’t appear in GET /v1/models — their pricing is flat and stated here and on each endpoint’s reference page: Charged on success only, like generation. POST /v1/cost also resolves them — send { "model": "convert-image" } (or convert-video / convert-audio / create-post). A post that is later hidden by content review is not refunded — it was published.

Estimating cost up front

To know what a request will cost before running it, POST the same body to /v1/cost with model added:
It’s computed by the same logic that bills you, so it never disagrees with the charge. exact is true for deterministic pricing; for seedream’s sequential mode (the model decides how many images, up to max_images) credits is the ceiling and exact is false — the real charge is always ≤ it.

Checking your balance

total_available is what you can spend right now (monthly remaining plus any purchased extra credits). If a request would exceed your available balance, it’s rejected before the job is created. For a multi-image request we reserve the maximum it could cost (max_images × credits_per_generation) up front, then charge only for the images actually returned.

Auto-reload

So a production integration doesn’t stop when credits run out, you can turn on auto-reload in the developer console under Billing: when your available balance (total_available above) drops below a threshold you choose, we charge a card saved on your account for one credit pack and add its credits, at the same price, plan bonus and expiry as buying the pack yourself.
  • You set the limits. A monthly limit in US dollars is required, and auto-reload runs at most 3 times in any 24 hours. Charges are in USD.
  • Confirmed by email. Turning it on or changing it needs a code sent to your account email, so a key or a connected app can’t turn it on for you. Turning it off is one click.
  • Receipts. Every charge sends a receipt email; the console shows recent activity.
  • Card requirements. The card must already have paid us at least 7 days ago. Cards issued in India can’t be charged automatically under bank mandate rules, so top up manually.
  • A declined charge pauses auto-reload and emails you; we don’t retry on our own.
The balance is checked about once a minute, so set a threshold that covers a few minutes of your heaviest usage.

Spend limits and alerts

You can put guardrails on API spend in the developer console under Billing. Both controls count programmatic traffic — API keys and the MCP connector — since both can run unattended. Usage in the web app is separate and never blocked by them.
  • Monthly spend limit — once your API usage reaches this many credits in a calendar month, new /v1 requests that cost credits and MCP generations are rejected until you raise or remove the limit, or the month rolls over. Jobs that were already accepted finish and bill normally, so a month can close slightly above the cap. Setting the limit to 0 pauses all API and MCP traffic immediately (a kill switch for a leaked key — though you should also revoke the key).
  • Spend alert — an email to your account address when the month’s API usage crosses a threshold. Purely informational, sent at most once per month; nothing is blocked. Set it below your limit to get warning before requests pause.
Both controls are off by default — credits are prepaid, so your balance already caps exposure, and a default hard cap could interrupt production traffic you intended to run. Thresholds are set in whole thousands of credits (1k credits ≈ $1 at pack rates), so a limit reads the same in credits and dollars. Requests over the limit receive HTTP 429 with:
Unlike rate_limited, backing off won’t clear this — don’t retry on monthly_spend_limit_reached; raise the limit (or wait for the new month) instead. The Retry-After header on this response is the number of seconds until the month rolls over. Reaching the limit only blocks requests that can spend credits. Job status (GET /v1/images/{id}, /v1/videos/{id}, /v1/audios/{id} and the other job endpoints), GET /v1/models, GET /v1/account/credits and POST /v1/cost keep working, so you can still collect jobs that started before the limit. The stock searches cost credits, so they are blocked too.