Nemotron 3.5 Super API: Pricing, Documentation

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/nemotron-3-5-super-text/v1/text-to-text' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"prompt": "Explain in two sentences why the sky looks blue.", "max_tokens": 256}'

Callback Payload (success)

{
  "request_id": "nemotron-3-5-super-text_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "nemotron-3-5-super-text",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json"
    ],
    "media_type": "application/json",
    "output": "The answer.",
    "finish_reason": "stop",
    "usage": {
      "input_tokens": 52,
      "output_tokens": 61
    }
  },
  "created_at": "2026-10-09T10:04:54.387Z",
  "completed_at": "2026-10-09 10:06:21"
}

Failure callback shape

{
  "request_id": "nemotron-3-5-super-text_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "nemotron-3-5-super-text",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request returns one answer. batch_size, num_images, number_of_images, n and count are rejected — send one request per answer.
• This API has no length setting: use max_tokens to cap the answer. duration, seconds, num_frames, frames_per_second and music_length_ms are rejected.
• Send every field at the top level of the JSON body; parameters, content and text are rejected.
• max_tokens takes a whole number from 1 to 32000 (1024 when you leave it out); any other value is refused at submit.
• image_url, video_url, image_urls and video_urls are rejected (this endpoint answers text only: use Nemotron 3.5 Super Vision for an image or Nemotron 3.5 Super Video for a video).
• messages is rejected (send the question as prompt).
• max_new_tokens and max_completion_tokens are rejected (use max_tokens).
• Use the JSON types shown: prompt, system_prompt and reasoning_mode are strings, max_tokens is a whole number (not "1024"), temperature and top_p are numbers, and enable_safety_checker is true or false.
• Omit a field instead of sending null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url, webhook_mode and callback_url in the JSON body are rejected.
• stream is not supported: the result is delivered when the job completes.
• Options this API does not offer are rejected: input and model.

Reasoning and max_tokens

• max_tokens caps everything the model generates, its reasoning included. When the cap is reached, output.finish_reason is length instead of stop.
• With reasoning_mode think, the reasoning comes first: a small max_tokens can be used up by the reasoning, and output.output is then an empty string. Allow a few thousand tokens for think.
• An empty or cut-short answer is billed like any other: you pay for the tokens used, shown in output.usage.

Example: show the reasoning

{
  "prompt": "Is 1001 a prime number? Answer yes or no.",
  "reasoning_mode": "think",
  "max_tokens": 4096
}

Instructions

• Put standing instructions (tone, length, format) in system_prompt and the question itself in prompt.
• There is no conversation history: each request stands alone. To continue a conversation, include the earlier turns in prompt.

Example: instructions in system_prompt

{
  "system_prompt": "You are a concise assistant. Answer in one sentence.",
  "prompt": "What does an API gateway do?",
  "temperature": 0.6
}

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing prompt (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' is required",
  "violations": [
    "'prompt' is required"
  ]
}

max_tokens above 32000 (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'max_tokens' must be <= 32000 (got 40000)",
  "violations": [
    "'max_tokens' must be <= 32000 (got 40000)"
  ]
}

A chat-style messages list (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' is required; 'messages' must be one of:  (got [{\"role\":\"user\",\"content\":\"Hello\"}]); 'messages' is not accepted. Send the question as 'prompt' and optional instructions as 'system_prompt'.",
  "violations": [
    "'prompt' is required",
    "'messages' must be one of:  (got [{\"role\":\"user\",\"content\":\"Hello\"}])",
    "'messages' is not accepted. Send the question as 'prompt' and optional instructions as 'system_prompt'."
  ]
}

An image sent to the text endpoint (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'image_url' must be one of:  (got \"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\"); 'image_url' is not accepted. This endpoint answers text only: use nemotron-3-5-super-vision for an image or nemotron-3-5-super-video for a video.",
  "violations": [
    "'image_url' must be one of:  (got \"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\")",
    "'image_url' is not accepted. This endpoint answers text only: use nemotron-3-5-super-vision for an image or nemotron-3-5-super-video for a video."
  ]
}

More than one answer per request (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'n' must be one of:  (got 2); 'n' is not supported: each request returns one answer. Send one request per question.",
  "violations": [
    "'n' must be one of:  (got 2)",
    "'n' is not supported: each request returns one answer. Send one request per question."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then read the answer in output.output (the same JSON is stored as a file at output.media_url[0]). Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/nemotron-3-5-super-text_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "nemotron-3-5-super-text_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "nemotron-3-5-super-text",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json"
    ],
    "media_type": "application/json",
    "output": "The answer.",
    "finish_reason": "stop",
    "usage": {
      "input_tokens": 52,
      "output_tokens": 61
    }
  },
  "created_at": "2026-10-09T10:04:54.387Z",
  "completed_at": "2026-10-09 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/nemotron-3-5-super-text/v1/text-to-text.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, read the answer in output.output (the same JSON is stored as a file at output.media_url[0]).

How Billing Is Measured

Billing is measured per 1,000 tokens of input plus output (your prompt and instructions count as input; the answer and any reasoning count as output), rounded up to the next 1,000, with a minimum of one unit per request. A hold for the most the request can use is placed when you submit; when the answer is ready, the charge is set from the tokens actually used, shown in output.usage, and the rest of the hold is released. A failed request is not charged. The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/nemotron-3-5-super-vision/v1/image-to-text' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"prompt": "Describe this image in one sentence.", "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg", "max_tokens": 256}'

Callback Payload (success)

{
  "request_id": "nemotron-3-5-super-vision_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "nemotron-3-5-super-vision",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json"
    ],
    "media_type": "application/json",
    "output": "The answer.",
    "finish_reason": "stop",
    "usage": {
      "input_tokens": 52,
      "output_tokens": 61
    }
  },
  "created_at": "2026-10-09T10:04:54.387Z",
  "completed_at": "2026-10-09 10:06:21"
}

Failure callback shape

{
  "request_id": "nemotron-3-5-super-vision_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "nemotron-3-5-super-vision",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request returns one answer. batch_size, num_images, number_of_images, n and count are rejected — send one request per answer.
• This API has no length setting: use max_tokens to cap the answer. duration, seconds, num_frames, frames_per_second and music_length_ms are rejected.
• Send every field at the top level of the JSON body; parameters, content and text are rejected.
• Send one image as image_url: a publicly reachable https URL (recommended) or a base64 data URI of a JPEG or PNG. An image in another format is refused at submit, with nothing charged.
• max_tokens takes a whole number from 1 to 32000 (1024 when you leave it out); any other value is refused at submit.
• video_url, video_urls and image_urls are rejected (this endpoint reads one image, as image_url; use Nemotron 3.5 Super Video for a video).
• messages is rejected (send the question as prompt).
• max_new_tokens and max_completion_tokens are rejected (use max_tokens).
• Use the JSON types shown: prompt, system_prompt and reasoning_mode are strings, max_tokens is a whole number (not "1024"), temperature and top_p are numbers, and enable_safety_checker is true or false.
• Omit a field instead of sending null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url, webhook_mode and callback_url in the JSON body are rejected.
• stream is not supported: the result is delivered when the job completes.
• Options this API does not offer are rejected: input and model.

Reasoning and max_tokens

• max_tokens caps everything the model generates, its reasoning included. When the cap is reached, output.finish_reason is length instead of stop.
• With reasoning_mode think, the reasoning comes first: a small max_tokens can be used up by the reasoning, and output.output is then an empty string. Allow a few thousand tokens for think.
• An empty or cut-short answer is billed like any other: you pay for the tokens used, shown in output.usage.
• The image counts as input tokens: a larger image uses more of them.

Example: show the reasoning

{
  "prompt": "How many people are in this image? Explain briefly.",
  "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg",
  "reasoning_mode": "think",
  "max_tokens": 4096
}

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing image_url (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'image_url' is required",
  "violations": [
    "'image_url' is required"
  ]
}

A reasoning mode that does not exist (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'reasoning_mode' must be one of: \"think\", \"no_think\" (got \"deep\")",
  "violations": [
    "'reasoning_mode' must be one of: \"think\", \"no_think\" (got \"deep\")"
  ]
}

A video sent to the image endpoint (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'video_url' must be one of:  (got \"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-5s.mp4\"); 'video_url' is not accepted. This endpoint reads one image: use nemotron-3-5-super-video for a video.",
  "violations": [
    "'video_url' must be one of:  (got \"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-5s.mp4\")",
    "'video_url' is not accepted. This endpoint reads one image: use nemotron-3-5-super-video for a video."
  ]
}

More than one image per request (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'image_urls' must be one of:  (got [\"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\",\"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\"]); 'image_urls' is not accepted: each request reads at most one image, as 'image_url' (nemotron-3-5-super-vision).",
  "violations": [
    "'image_urls' must be one of:  (got [\"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\",\"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\"])",
    "'image_urls' is not accepted: each request reads at most one image, as 'image_url' (nemotron-3-5-super-vision)."
  ]
}

top_p of 0 (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'top_p' must be >= 0.01 (got 0)",
  "violations": [
    "'top_p' must be >= 0.01 (got 0)"
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then read the answer in output.output (the same JSON is stored as a file at output.media_url[0]). Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/nemotron-3-5-super-vision_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "nemotron-3-5-super-vision_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "nemotron-3-5-super-vision",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json"
    ],
    "media_type": "application/json",
    "output": "The answer.",
    "finish_reason": "stop",
    "usage": {
      "input_tokens": 52,
      "output_tokens": 61
    }
  },
  "created_at": "2026-10-09T10:04:54.387Z",
  "completed_at": "2026-10-09 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/nemotron-3-5-super-vision/v1/image-to-text.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, read the answer in output.output (the same JSON is stored as a file at output.media_url[0]).

How Billing Is Measured

Billing is measured per 1,000 tokens of input plus output (your prompt, instructions and the image count as input; the answer and any reasoning count as output), rounded up to the next 1,000, with a minimum of one unit per request. A hold for the most the request can use is placed when you submit; when the answer is ready, the charge is set from the tokens actually used, shown in output.usage, and the rest of the hold is released. A failed request is not charged. The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/nemotron-3-5-super-video/v1/video-to-text' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"prompt": "Summarize what happens in this video in one sentence.", "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-5s.mp4", "max_tokens": 256}'

Callback Payload (success)

{
  "request_id": "nemotron-3-5-super-video_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "nemotron-3-5-super-video",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json"
    ],
    "media_type": "application/json",
    "output": "The answer.",
    "finish_reason": "stop",
    "usage": {
      "input_tokens": 52,
      "output_tokens": 61
    }
  },
  "created_at": "2026-10-09T10:04:54.387Z",
  "completed_at": "2026-10-09 10:06:21"
}

Failure callback shape

{
  "request_id": "nemotron-3-5-super-video_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "nemotron-3-5-super-video",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request returns one answer. batch_size, num_images, number_of_images, n and count are rejected — send one request per answer.
• This API has no length setting: use max_tokens to cap the answer. duration, seconds, num_frames, frames_per_second and music_length_ms are rejected.
• Send every field at the top level of the JSON body; parameters, content and text are rejected.
• Send one video as video_url: a publicly reachable https URL (recommended) or a base64 data URI of an MP4 of 2 minutes or less. A video measured longer than 2 minutes is refused at submit, with nothing charged.
• max_tokens takes a whole number from 1 to 32000 (1024 when you leave it out); any other value is refused at submit.
• image_url, image_urls and video_urls are rejected (this endpoint reads one video, as video_url; use Nemotron 3.5 Super Vision for an image).
• messages is rejected (send the question as prompt).
• max_new_tokens and max_completion_tokens are rejected (use max_tokens).
• Use the JSON types shown: prompt, system_prompt and reasoning_mode are strings, max_tokens is a whole number (not "1024"), temperature and top_p are numbers, and enable_safety_checker is true or false.
• Omit a field instead of sending null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url, webhook_mode and callback_url in the JSON body are rejected.
• stream is not supported: the result is delivered when the job completes.
• Options this API does not offer are rejected: input and model.

Reasoning and max_tokens

• max_tokens caps everything the model generates, its reasoning included. When the cap is reached, output.finish_reason is length instead of stop.
• With reasoning_mode think, the reasoning comes first: a small max_tokens can be used up by the reasoning, and output.output is then an empty string. Allow a few thousand tokens for think.
• An empty or cut-short answer is billed like any other: you pay for the tokens used, shown in output.usage.
• The video counts as input tokens: a longer video uses more of them. The content of the frames does not change the count.

Example: show the reasoning

{
  "prompt": "What changes between the start and the end of this video?",
  "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-5s.mp4",
  "reasoning_mode": "think",
  "max_tokens": 4096
}

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing video_url (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'video_url' is required",
  "violations": [
    "'video_url' is required"
  ]
}

A length setting (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'duration' must be one of:  (got 10); 'duration' is not accepted: this API has no length setting. Use 'max_tokens' to cap the answer.",
  "violations": [
    "'duration' must be one of:  (got 10)",
    "'duration' is not accepted: this API has no length setting. Use 'max_tokens' to cap the answer."
  ]
}

An image sent to the video endpoint (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'image_url' must be one of:  (got \"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\"); 'image_url' is not accepted. This endpoint reads one video: use nemotron-3-5-super-vision for an image.",
  "violations": [
    "'image_url' must be one of:  (got \"https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg\")",
    "'image_url' is not accepted. This endpoint reads one video: use nemotron-3-5-super-vision for an image."
  ]
}

max_tokens as a string (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'max_tokens' must be integer (got string)",
  "violations": [
    "'max_tokens' must be integer (got string)"
  ]
}

Instructions inside an input object (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' is required; 'input' must be one of:  (got {\"prompt\":\"Summarize what happens in this video in one sentence.\"}); 'input' is not accepted by this model. Send the fields at the top level of the JSON body.",
  "violations": [
    "'prompt' is required",
    "'input' must be one of:  (got {\"prompt\":\"Summarize what happens in this video in one sentence.\"})",
    "'input' is not accepted by this model. Send the fields at the top level of the JSON body."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then read the answer in output.output (the same JSON is stored as a file at output.media_url[0]). Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/nemotron-3-5-super-video_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "nemotron-3-5-super-video_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "nemotron-3-5-super-video",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json"
    ],
    "media_type": "application/json",
    "output": "The answer.",
    "finish_reason": "stop",
    "usage": {
      "input_tokens": 52,
      "output_tokens": 61
    }
  },
  "created_at": "2026-10-09T10:04:54.387Z",
  "completed_at": "2026-10-09 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/nemotron-3-5-super-video/v1/video-to-text.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, read the answer in output.output (the same JSON is stored as a file at output.media_url[0]).

How Billing Is Measured

Billing is measured per 1,000 tokens of input plus output (your prompt, instructions and the video count as input; the answer and any reasoning count as output), rounded up to the next 1,000, with a minimum of one unit per request. A hold for the most the request can use is placed when you submit; when the answer is ready, the charge is set from the tokens actually used, shown in output.usage, and the rest of the hold is released. A failed request is not charged. The current rate is shown in the pricing panel on this page.