inference.minimax.h3.hyper.txt2vid.v0

The inference.minimax.h3.hyper.txt2vid.v0 job generates a video with synchronized stereo audio (1344×768, 24 fps, AAC 32 kHz) from a text prompt using the MiniMax H3 (Hailuo-03) model. The hyper tier is an 8-evaluation speed tier in the same latency class as turbo; outputs are equally coherent but not pixel-identical to base, fast or turbo at the same seed.

{
    "type": "inference.minimax.h3.hyper.txt2vid.v0",
    "config": {
        "prompt": "a red fox trotting through a snowy pine forest at golden hour"
    }
}

Prompting

The prompt is used exactly as written: nothing rewrites or expands it before it reaches the model, which was trained on descriptive captions. Describe the shot rather than making a request: the subject, setting, time of day, weather, lighting, camera move and motion. For example, “show me what christmas would be like in Tokyo” tends to produce a generic winter scene, while “Shibuya crossing in Tokyo on Christmas Eve at night, light snow falling, crowds with umbrellas, a giant illuminated Christmas tree, neon signs reflecting in the wet street, slow cinematic dolly shot” produces the intended scene. The hyper tier is a distilled 8-evaluation recipe, so it follows the prompt less closely than base and fast; vague prompts steer it even less, so a detailed description matters more here.

duration (whole seconds, 4–15), aspect_ratio (one of 16:9, 4:3, 1:1, 3:4, 9:16, 21:9; default 16:9) and seed can be set. Durations round up to the model’s frame grid (17n+5 frames at 24 fps): 4 s returns 4.46 s, 15 s returns 15.08 s, and values outside 4–15 fail validation. The same seed reproduces the same video within a tier. resolution is accepted for parity with other MiniMax job types, but only its default 768P is valid: this job type uses the 768P canvas preset, targeting a 768 px short edge with a 768×1344 pixel area cap before rounding to multiples of 32.

Typical latency: roughly 21 s for a 10 s clip, ~38 s for a 15 s one.

Schema

{
  "type": "object",
  "required": [
    "type",
    "config"
  ],
  "additionalProperties": false,
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "inference.minimax.h3.base.txt2vid.v0",
        "inference.minimax.h3.fast.txt2vid.v0",
        "inference.minimax.h3.base.txt2vid.v1",
        "inference.minimax.h3.fast.txt2vid.v1",
        "inference.minimax.h3.turbo.txt2vid.v0",
        "inference.minimax.h3.hyper.txt2vid.v0",
        "inference.minimax.h3.chili.txt2vid.v0"
      ]
    },
    "config": {
      "type": "object",
      "required": [
        "prompt"
      ],
      "additionalProperties": false,
      "properties": {
        "prompt": {
          "type": "string",
          "minLength": 3,
          "maxLength": 7000,
          "default": "a dancing cat under moonlight",
          "description": "Description of the desired video and audio."
        },
        "duration": {
          "type": "integer",
          "minimum": 4,
          "maximum": 15,
          "default": 6,
          "description": "Duration of the video in seconds. Rounded up to the model's frame grid (4 s returns 4.46 s, 15 s returns 15.08 s)."
        },
        "aspect_ratio": {
          "type": "string",
          "enum": [
            "16:9",
            "4:3",
            "1:1",
            "3:4",
            "9:16",
            "21:9"
          ],
          "default": "16:9",
          "description": "Controls the approximate aspect ratio of the resulting video."
        },
        "resolution": {
          "type": "string",
          "enum": [
            "768P"
          ],
          "default": "768P",
          "description": "Only 768P is supported: the canvas targets a 768px short edge with a 768x1344 pixel area cap and 32px rounding; exact width and height follow the selected aspect ratio."
        },
        "seed": {
          "type": "integer",
          "minimum": 0,
          "maximum": 2147483647,
          "description": "Seed for reproducible generation. Random when omitted."
        }
      }
    }
  }
}