# Bytedance Seed V2 Mini API

> LLM multimodal que lee texto, imágenes y video, con razonamiento configurable e historial de varios turnos

Modelo de lenguaje que responde a un `prompt` de texto y también puede analizar hasta 6 imágenes (`image_urls`) y hasta 3 videos (`video_urls`), sin comprensión de audio. `thinking` puede estar activado, desactivado o en auto, y `reasoning_effort` va de minimal a high. La longitud de salida llega hasta 65536 tokens mediante `max_completion_tokens`. Admite un `system_prompt` y `messages` anteriores para conversaciones de varios turnos.

- Model id: `bytedance/seed/v2/mini`
- Made by: ByteDance
- Category: LLM
- Output: text
- Charged by usage (up to $0.213 per run, typical $0.001; the rest of the hold is refunded): $0.0001 / 1000 tokens (cheapest run $0.001 (1 API credit))
- Run: `POST https://api.doitong.com/v1/run/bytedance/seed/v2/mini`
- Quote: `POST https://api.doitong.com/v1/quote/bytedance/seed/v2/mini`
- Docs page: https://doitong.com/es/developers/models/bytedance/seed/v2/mini

## Pricing

Price: **$0.0001 / 1000 tokens**. Charged by actual usage at $0.0001 / 1000 tokens: up to $0.213 per run (the quote gives the maximum for your input), the rest is refunded automatically after the run. A typical run costs $0.001.

The full example input below holds up to $0.213 = 213 API credits (up to 2129.92 × 1000 tokens × $0.0001) when it starts; a typical run costs $0.001 (1 API credit).

Prices are the upstream price in USD, paid from the API balance in API credits (1000 API credits = 1 USD). This model is charged by usage: `POST https://api.doitong.com/v1/quote/bytedance/seed/v2/mini` with the same JSON body returns the most the run can cost for that input (`credits`, debited when the run starts); after the run you pay the usage the model reports and the rest is refunded automatically. An input whose maximum would exceed the per-run limit is refused with 422. Invalid input is never charged and failed or canceled requests are refunded automatically.

## Input parameters

Send either the canonical name or the native field name (not both). Canonical names are the same across models.

| Parameter | Canonical name | Type | Required | Default | Allowed values / range | Description |
|---|---|---|---|---|---|---|
| `prompt` |  | string | yes |  |  | Texto que contiene el prompt o la pregunta enviada al modelo. |
| `image_urls` | `images` | array of image URL | no |  |  | Proporciona imágenes para la comprensión visual mediante URL que apunten a archivos JPEG, PNG o WebP. El procesamiento está limitado a 6 imágenes; las entradas que superen ese límite se ignoran. |
| `video_urls` | `videos` | array of video URL | no |  |  | Proporciona URL de videos MP4 o MOV para comprender su contenido visual, no su audio. Solo se pueden procesar 3 videos; las entradas que superen ese límite se ignoran. |
| `system_prompt` |  | string \| null | no |  |  | Prompt opcional a nivel de sistema que dirige el comportamiento del modelo. |
| `messages` |  | array of object {role, content} \| null | no |  |  | Contexto opcional de varios turnos proporcionado mediante la reutilización del valor `messages` de una respuesta anterior. Independientemente de este historial, siempre se añade un nuevo turno de usuario final que contiene el `prompt`, `image_urls`, `video_urls` y `system_prompt` actuales. |
| `messages[].role` |  | enum (string) | yes |  | `system`, `user`, `assistant` | Rol asignado al autor del mensaje. |
| `messages[].content` |  | string \| array of object | yes |  |  | El contenido del mensaje acepta cualquiera de estas representaciones: una lista de partes para contenido multimodal, como imágenes, o una cadena cuando el mensaje solo contiene texto. |
| `thinking` |  | enum (string) | no | `enabled` | `enabled`, `disabled`, `auto` | Modo de razonamiento de cadena de pensamiento: `enabled` incluye el razonamiento cada vez, `disabled` lo excluye por completo y `auto` permite que el modelo elija según la consulta. |
| `reasoning_effort` |  | enum (string) \| null | no |  | `minimal`, `low`, `medium`, `high` | Elige con qué profundidad razona el modelo antes de responder: `high` analiza profundamente los problemas complejos, `medium` equilibra profundidad y velocidad, `low` usa un razonamiento ligero para responder más rápido y `minimal` responde inmediatamente. Esta configuración solo se aplica si `thinking` tiene el valor `enabled` o `auto`. |
| `max_completion_tokens` |  | integer | no | `4096` | 1 – 65,536 | Límite máximo de token para la salida generada. Tanto los tokens de respuesta como los tokens de cadena de pensamiento cuentan para este límite compartido. |
| `temperature` |  | number | no | `1` | 0 – 2 | Configuración de aleatoriedad de la respuesta que orienta la generación hacia la creatividad al aumentar o hacia una salida enfocada y determinista al disminuir. |
| `top_p` |  | number | no | `0.7` | 0 – 1 | El control de muestreo de núcleo `top_p` limita más estrechamente la selección de tokens al disminuir. Los tokens candidatos del modelo cubren una masa de probabilidad acumulada de `top_p`. |
| `enable_safety_checker` |  | boolean | no | `true` |  | La moderación puede relajarse estableciendo false, siempre que el autor de la llamada esté autorizado. |

File parameters (`images`, `videos`) take a public `https://` URL, a `data:` URI up to 10 MB, or the `file_url` of an upload from `POST /files`.

## Example input

Required parameters only:

```json
{
  "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field"
}
```

A fuller example:

```json
{
  "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field"
}
```

## Run it

Set `DOITONG_API_KEY` to a key from https://doitong.com/developers/dashboard#keys. `?wait=20` returns the finished request in the same call when it is ready within 20 seconds; otherwise poll `GET /requests/{id}` (it also accepts `?wait=`), or pass `webhook_url`.

### curl

```bash
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://api.doitong.com/v1/run/bytedance/seed/v2/mini?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field"
  }'

# Still "queued" or "running"? Poll with the id from the response:
curl "https://api.doitong.com/v1/requests/REQUEST_ID?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY"
```

### JavaScript

```javascript
const API = "https://api.doitong.com/v1";
const headers = {
  "Authorization": "Bearer " + process.env.DOITONG_API_KEY,
  "Content-Type": "application/json",
};

// Run the model; wait up to 20 s for the finished result in the same call
let res = await fetch(API + "/run/bytedance/seed/v2/mini?wait=20", {
  method: "POST",
  headers,
  body: JSON.stringify({
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field"
  }),
});
let request = await res.json();
if (!res.ok) throw new Error(request.error.code + ": " + request.error.message);

// Not finished yet: long-poll until it is
while (request.status === "queued" || request.status === "running") {
  res = await fetch(API + "/requests/" + request.id + "?wait=20", { headers });
  request = await res.json();
}
if (request.status !== "succeeded") throw new Error(JSON.stringify(request.error));
console.log(request.outputs.map((o) => o.url), request.cost.credits + " API credits ($" + request.cost.usd + ")");
```

### Python

```python
import os
import requests

API = "https://api.doitong.com/v1"
headers = {"Authorization": "Bearer " + os.environ["DOITONG_API_KEY"]}

payload = {
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
}

# Run the model; wait up to 20 s for the finished result in the same call
r = requests.post(API + "/run/bytedance/seed/v2/mini", params={"wait": 20}, json=payload, headers=headers)
request = r.json()
if not r.ok:
    raise RuntimeError(request["error"]["code"] + ": " + request["error"]["message"])

# Not finished yet: long-poll until it is
while request["status"] in ("queued", "running"):
    request = requests.get(API + "/requests/" + request["id"], params={"wait": 20}, headers=headers).json()
if request["status"] != "succeeded":
    raise RuntimeError(request["error"])
print([o["url"] for o in request["outputs"]], request["cost"]["credits"], "API credits", "$" + request["cost"]["usd"])
```

### MCP

The MCP server at `https://api.doitong.com/v1/mcp` exposes the tools `search_models`, `get_model`, `quote`, `run`, `get_request`, `cancel_request`, `list_requests`, `create_upload` and `get_account`. A `run` call for this model:

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "run",
    "arguments": {
      "model_id": "bytedance/seed/v2/mini",
      "input": {
        "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field"
      },
      "wait_seconds": 20
    }
  }
}
```

## Response

Every call returns a request object. `status` is one of `queued`, `running`, `succeeded`, `failed`, `canceled`. `outputs` lists the generated files with URLs on our storage (kept forever, or 7 days with `?storage=temporary`; `storage` says which); `output` is the model's own result object with the same URLs.

```json
{
  "id": "req_8k2m4q7x1z9c3v5b6n0p2r4t",
  "object": "request",
  "model": "bytedance/seed/v2/mini",
  "status": "succeeded",
  "created_at": "2026-10-20T12:00:00Z",
  "started_at": "2026-10-20T12:00:02Z",
  "completed_at": "2026-10-20T12:00:09Z",
  "queue_position": null,
  "cost": {
    "credits": 213,
    "usd": "0.213",
    "charged": 213,
    "refunded": 0
  },
  "input": {
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field"
  },
  "output": {
    "output": "string",
    "reasoning_content": "string",
    "messages": [
      {
        "role": "system",
        "content": "string"
      }
    ]
  },
  "outputs": [
    {
      "type": "file",
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
      "content_type": "image/jpeg",
      "file_size": 1843200,
      "stored": true
    }
  ],
  "error": null,
  "metrics": {
    "inference_seconds": 6.8
  },
  "urls": {
    "self": "https://api.doitong.com/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t",
    "cancel": "https://api.doitong.com/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t/cancel"
  }
}
```

### Output fields (`output`)

| Field | Type | Description |
|---|---|---|
| `output` | string | Texto devuelto por el modelo. |
| `reasoning_content` | string \| null | Con `thinking` establecido en `enabled` o `auto`, esta salida puede contener el razonamiento de cadena de pensamiento del modelo; está ausente con las demás configuraciones. |
| `messages` | array of object {role, content} | Contexto de conversación reutilizable para la siguiente entrada `messages`: el historial devuelto contiene todos los turnos, incluida la respuesta más reciente del modelo. |

## Related models

- [Fibo Edit \[Structured Instruction\]](https://doitong.com/developers/models/bria/fibo-edit/edit/structured_instruction.md) (`bria/fibo-edit/edit/structured_instruction`): $0.01 / request
- [Fibo Lite · Structured Prompt](https://doitong.com/developers/models/bria/fibo-lite/generate/structured_prompt.md) (`bria/fibo-lite/generate/structured_prompt`): $0.008 / request
- [Fibo · Structured Prompt](https://doitong.com/developers/models/bria/fibo/generate/structured_prompt.md) (`bria/fibo/generate/structured_prompt`): $0.01 / request
- [Mureka 9.5 Lyrics Generation](https://doitong.com/developers/models/mureka/api/generate/lyrics.md) (`mureka/api/generate/lyrics`): $0.0135 / request
- [Nemotron 3 Nano Omni](https://doitong.com/developers/models/nvidia/nemotron-3-nano-omni.md) (`nvidia/nemotron-3-nano-omni`): $0.01 / run
- [NVIDIA Nemotron 3.5 Super Text](https://doitong.com/developers/models/nvidia/nemotron-3-5-super/text.md) (`nvidia/nemotron-3-5-super/text`): $0.01 / run
- [OpenRouter](https://doitong.com/developers/models/openrouter/router.md) (`openrouter/router`): $0.01 / usage unit
- [OpenRouter \[Audio\]](https://doitong.com/developers/models/openrouter/router/audio.md) (`openrouter/router/audio`): $0.01 / usage unit

## Links

- HTML page: https://doitong.com/developers/models/bytedance/seed/v2/mini
- All models and guides for AI agents: https://doitong.com/developers/llms.txt
- OpenAPI 3.1 with this model's typed request body: https://doitong.com/developers/openapi.json?model=bytedance/seed/v2/mini
- Errors: https://doitong.com/developers/docs/errors
- Model details as JSON: https://api.doitong.com/v1/models/bytedance/seed/v2/mini