Minimax H3 Text To Image API Documentation

Minimax H3 Text To Image API Documentation

Playground

MiniMax H3 Open Weights Text-to-Image generates photorealistic, cinema-grade stills from a text prompt at 1K, 1.5K or 2K resolution with fifteen aspect ratios. Ready-to-use REST inference API, best performance, no cold starts, affordable pricing.

Features

Run the open-weights edition of MiniMax H3 as a still-image generator on WaveSpeedAI’s own GPU infrastructure. MiniMax H3 is a cinema-grade omni-modal generation model; this endpoint taps the same generator to produce a single photorealistic image from a text prompt, natively at 1k or 2k resolution.

Why Choose This?

• Cinematic realism
  Film-quality lighting, skin, fabric and depth of field from a video-grade generator, with strong prompt adherence.

• Native 2K output
  Generate up to ~4 megapixels directly, no separate upscaling step.

• Flexible aspect ratios
  Fifteen presets from square to tall, wide and cinematic formats.

• Fast and affordable
  Distilled sampling keeps generation to seconds per image at a flat per-image price.

Parameters

How to Use

1. Write your prompt — Describe the subject, setting, lighting, lens and mood as you would brief a cinematographer.
2. Pick an aspect ratio — Match the layout of your target format.
3. Choose resolution — 1k for fast iteration, 1.5k for a balance of detail and cost, 2k for final assets.
4. Submit — Receive the generated image URL.

Pricing

Pricing is based on the selected resolution.

Best Use Cases

• Portraits and lifestyle — Photorealistic people, expressions and skin detail.
• Key art and thumbnails — Cinematic stills that match your video content.
• Concept exploration — Iterate on scenes, costumes and moods before generating video.
• Marketing visuals — Campaign images, product scenes and social content.

Pro Tips

• Prompt like a film still: subject, action, environment, lighting, lens (e.g. 85mm, shallow depth of field), time of day.
• Use 2k for prints and hero images; 1k is usually enough for social and previews.
• Fix the seed to compare prompt variations on the same composition.

Related Models

• MiniMax H3 Image Edit — Re-render a subject from reference images into a new scene or style.
• MiniMax H3 Text-to-Video — Generate video with native audio from the same model family.

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "prompt": "A cinematic ocean wave at sunrise, highly detailed",
  "aspect_ratio": "1:1",
  "resolution": "1k",
  "output_format": "jpeg"
}
JSON
)

# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --connect-timeout 10 --max-time 60 \
  -X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/minimax-h3/text-to-image" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}"); then
  printf 'Submit failed. %s\n' "${CHECK}" >&2
  exit 1
fi

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id. %s\n' "${CHECK}" >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
  # Transient errors (429, 5xx) are retried, honoring Retry-After.
  HTTP_CODE=$(curl --silent --show-error --fail-with-body \
    --connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
    --output "${BODY_FILE}" --write-out '%{http_code}' \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
  case "${HTTP_CODE}" in
    2??) ;;
    000|429|5??)  # dropped connection or temporary error: keep polling
      printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      sleep 10; continue ;;
    *) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      cat "${BODY_FILE}" >&2; exit 1 ;;
  esac
  RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted)
      printf 'Task %s ended with status %s\n' "${PREDICTION_ID}" "${STATUS}" >&2
      printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Every model also accepts the global request options enable_sync_mode and enable_base64_output; see their pages for details.

Response Parameters

Result Request Parameters

Result Response Parameters