MODEL ID runware:ace-step@v1.5-xl-sft
live

ACE-Step v1.5 XL SFT

ACE-Step v1.5 XL SFT is the supervised fine-tuned 4B DiT variant in the ACE-Step 1.5 XL line. It is positioned as the highest-quality XL option, combining 50-step CFG inference with stronger prompt adherence and refined audio quality for text-to-music, cover, and repaint workflows when final output quality matters more than speed or broader editing task coverage.

ACE-Step v1.5 XL SFT
Text to Audio

Polar Freight Yard Hymn

0:00
{
  "taskType": "audioInference",
  "taskUUID": "c9fe7e9b-b1ea-426e-91c1-0d1f117c7518",
  "model": "runware:ace-step@v1.5-xl-sft",
  "positivePrompt": "A cinematic Arctic industrial folk-electronica track set around a remote polar freight yard at 3 AM: low throat-singing style male choir hums without words, bowed metal drones, pizzicato cello ostinato, granular snow-step percussion, deep sub-kick every other beat, distant rail brakes screeching as tonal accents, warm analog pad chords, steady 6/4 pulse, rising tension then serene release. High fidelity, spacious stereo image, crisp transient detail, realistic acoustic instruments blended with modern bass design, no lyrical vocals.",
  "negativePrompt": "muddy mix, harsh clipping, static noise, spoken narration, pop hook, generic EDM buildup, excessive reverb wash, off key vocals, distorted master, low bitrate artifacts, thin bass, chaotic timing",
  "duration": 95,
  "seed": 28003,
  "steps": 50,
  "CFGScale": 12,
  "settings": {
    "bpm": 84,
    "keyScale": "D minor",
    "timeSignature": 6,
    "guidanceType": "apg",
    "cfgIntervalStart": 0,
    "cfgIntervalEnd": 0.82,
    "vocalLanguage": "unknown"
  }
}
{
  "taskType": "audioInference",
  "taskUUID": "c9fe7e9b-b1ea-426e-91c1-0d1f117c7518",
  "audioUUID": "1edef3ab-824f-44d9-9360-00da747f1215",
  "audioURL": "https://am.runware.ai/audio/os/a08dlim3/ws/5/ai/1edef3ab-824f-44d9-9360-00da747f1215.mp3",
  "seed": 28003,
  "cost": 0.00765
}
Text to Audio

Volcanic Apiary Koto Breakbeat

0:00
{
  "taskType": "audioInference",
  "taskUUID": "174e0770-04d1-4f64-9c17-a5d0f749f2c9",
  "model": "runware:ace-step@v1.5-xl-sft",
  "positivePrompt": "A high-fidelity hybrid electronic song set around an imagined volcanic apiary: nimble koto plucks, tight taiko hits, rapid jungle breakbeats, warm analog bass, granular synth buzz shaped like swarming bees, and soaring female Japanese vocals. Start with sparse koto and distant hive-like drones, build into a crisp 174 BPM breakbeat drop, add call-and-response vocal layers in the chorus, then finish with a cinematic final refrain and clean sub-bass impact. Detailed mix, wide stereo field, punchy drums, expressive vocal phrasing, modern production, dramatic but elegant mood.",
  "negativePrompt": "muddy mix, distorted vocals, random speech, harsh clipping, off-key singing, weak drums, excessive reverb, low fidelity, flat arrangement, abrupt ending, background chatter",
  "duration": 78,
  "seed": 99824,
  "steps": 50,
  "CFGScale": 15,
  "settings": {
    "bpm": 174,
    "keyScale": "F# minor",
    "timeSignature": 4,
    "guidanceType": "cfg",
    "cfgIntervalStart": 0,
    "cfgIntervalEnd": 1,
    "vocalLanguage": "ja",
    "lyrics": "[Verse 1]\n黒い石の庭で\n小さな羽が鳴る\n熱い息を抱いて\n金の道を探す\n\n[Pre-Chorus]\n針のような星へ\n声を重ねて\n鼓動が速くなる\nまだ遠くへ\n\n[Chorus]\n蜂の夢よ 舞い上がれ\n火の花を越えて\n甘い影を 連れて行け\n夜の底まで\n\n[Bridge]\n琴の糸が光る\n太鼓が胸を打つ\nひとつ ふたつ 群れは\n空へほどける\n\n[Final Chorus]\n蜂の夢よ 舞い上がれ\n火の花を越えて\n甘い影を 連れて行け\n夜の底まで"
  }
}
{
  "taskType": "audioInference",
  "taskUUID": "174e0770-04d1-4f64-9c17-a5d0f749f2c9",
  "audioUUID": "9e287982-2905-460f-9e8d-2abc1232b7c9",
  "audioURL": "https://am.runware.ai/audio/os/a08dlim3/ws/5/ai/9e287982-2905-460f-9e8d-2abc1232b7c9.mp3",
  "seed": 99824,
  "cost": 0.0096
}
Text to Audio

Cedar Wetland Tabla Hyperpop

0:00
{
  "taskType": "audioInference",
  "taskUUID": "d18b46ed-c67a-44bc-9997-d9d63d56a349",
  "model": "runware:ace-step@v1.5-xl-sft",
  "positivePrompt": "High-fidelity 108 second song blending tabla-driven hyperpop, cedar flute motifs, rubbery synth bass, chopped vocal harmonies, marimba accents, tight electronic drums, and warm bass clarinet countermelodies. Mood is bright, curious, and slightly mischievous, like a hidden radio station broadcasting from a green wetland boardwalk at sunrise. Structure: short textured intro, punchy verse, soaring chorus, instrumental breakdown with tabla rolls and glitchy vocal cuts, final bigger chorus. Clean modern mix, crisp transients, wide stereo image, detailed low end, expressive lead vocal, catchy hook.",
  "negativePrompt": "lo-fi distortion, muddy mix, harsh clipping, off-key vocals, weak drums, excessive reverb, monotone singing, generic EDM drop, noisy artifacts, spoken narration, crowd noise",
  "duration": 108,
  "seed": 58656,
  "steps": 50,
  "CFGScale": 12,
  "settings": {
    "bpm": 142,
    "keyScale": "D minor",
    "timeSignature": 4,
    "guidanceType": "cfg",
    "vocalLanguage": "en",
    "cfgIntervalStart": 0,
    "cfgIntervalEnd": 1,
    "lyrics": "[Verse 1]\nFootsteps tap on cedar planks\nSilver reeds keep counting sparks\nPocket songs in paper maps\nFind the beat between our hearts\n\n[Pre-Chorus]\nIf the path bends, let it bend\nEvery echo finds a friend\n\n[Chorus]\nTurn the signal up, up, up\nWe are small but loud enough\nHandclap thunder, tabla rush\nCarry us, carry us\n\n[Verse 2]\nBassline curls through waking green\nFlutes reply in bright machine\nEvery note a secret key\nEvery breath a new routine\n\n[Bridge]\nCut the voices into light\nSpin the rhythm left and right\nHold the pulse and let it climb\nOne more spark inside the wire\n\n[Final Chorus]\nTurn the signal up, up, up\nWe are small but loud enough\nHandclap thunder, tabla rush\nCarry us, carry us"
  }
}
{
  "taskType": "audioInference",
  "taskUUID": "d18b46ed-c67a-44bc-9997-d9d63d56a349",
  "audioUUID": "eca53984-92dd-450f-99c1-2fe16cca08e5",
  "audioURL": "https://am.runware.ai/audio/os/a06dlim3/ws/5/ai/eca53984-92dd-450f-99c1-2fe16cca08e5.mp3",
  "seed": 58656,
  "cost": 0.01155
}
Text to Audio

Sulfur Cathedral Bass Procession

0:00
{
  "taskType": "audioInference",
  "taskUUID": "780d127e-d0e0-4a7c-b460-f69e211d1cd8",
  "model": "runware:ace-step@v1.5-xl-sft",
  "positivePrompt": "A dark cinematic trip-hop hymn set inside a vast sulfur crystal cathedral, slow 6/8 pulse, deep sub bass, brushed frame drums, bowed metal plates, low male lead vocal with distant female choir responses, granular synth drones, sparse piano notes, ritual procession energy, intimate verses expanding into a huge final chorus, pristine studio mix, wide stereo space, dramatic but controlled dynamics, textured organic percussion, film-score tension blended with alternative soul.",
  "negativePrompt": "muddy mix, harsh clipping, off-key vocals, thin bass, random noise, speech fragments, crowd sounds, abrupt ending, messy rhythm, low fidelity, excessive reverb, distorted lead vocal",
  "duration": 96,
  "seed": 7322,
  "steps": 50,
  "CFGScale": 14.5,
  "settings": {
    "bpm": 86,
    "timeSignature": 6,
    "keyScale": "D minor",
    "guidanceType": "cfg",
    "cfgIntervalStart": 0,
    "cfgIntervalEnd": 1,
    "vocalLanguage": "en",
    "lyrics": "[Verse 1]\nStone ribs hum beneath my feet\nAsh on my hands, a vow to keep\nEvery echo knows my name\nEvery spark returns the flame\n\n[Pre-Chorus]\nCarry the weight, carry the sound\nDown where the old roots coil around\n\n[Chorus]\nWe move as one through the burning hall\nLow drums answer when the shadows call\nRaise your voice where the crystals sing\nFrom broken dust, we become the ring\n\n[Verse 2]\nIron bells in the cavern throat\nMark each step with a solemn note\nNo crown, no throne, no golden sign\nOnly breath and the bassline\n\n[Final Chorus]\nWe move as one through the burning hall\nLow drums answer when the shadows call\nRaise your voice where the crystals sing\nFrom broken dust, we become the ring"
  }
}
{
  "taskType": "audioInference",
  "taskUUID": "780d127e-d0e0-4a7c-b460-f69e211d1cd8",
  "audioUUID": "fdee1533-0261-427b-a2b2-e7ef9e8c938e",
  "audioURL": "https://am.runware.ai/audio/os/a08dlim3/ws/5/ai/fdee1533-0261-427b-a2b2-e7ef9e8c938e.mp3",
  "seed": 7322,
  "cost": 0.00675
}
Text to Audio

Mossy Subway Choir Garage

0:00
{
  "taskType": "audioInference",
  "taskUUID": "3431b723-fc44-450a-9ac7-2a550c58539a",
  "model": "runware:ace-step@v1.5-xl-sft",
  "positivePrompt": "High-quality vocal UK garage track set in an abandoned underground rail concourse overtaken by moss, ceramic tiles, and distant electrical hum. 132 BPM, D minor, tight shuffled drums, warm sub bass, crisp rim clicks, chopped vocal echoes, soft organ chords, plucked harp accents, and a small choir answering the lead singer in short phrases. The mix should feel spacious, clean, cinematic, and emotionally hopeful, with a strong verse, lifted pre-chorus, memorable chorus, brief instrumental bridge, and final chorus with layered harmonies.",
  "negativePrompt": "muddy low end, harsh clipping, off-key vocals, weak drums, excessive reverb, noisy artifacts, flat arrangement, inaudible lyrics, chaotic percussion, distorted master",
  "duration": 75,
  "seed": 39665,
  "steps": 50,
  "CFGScale": 15,
  "settings": {
    "bpm": 132,
    "keyScale": "D minor",
    "timeSignature": 4,
    "vocalLanguage": "en",
    "guidanceType": "cfg",
    "cfgIntervalStart": 0,
    "cfgIntervalEnd": 1,
    "lyrics": "[Verse]\nUnder the rails where the old signs glow\nFootsteps fade where the green roots grow\nI found your name in a broken tone\nSinging back through the stone alone\n\n[Pre-Chorus]\nIf the power cuts, we carry the spark\nHandclaps rising from the dark\n\n[Chorus]\nMeet me down where the echoes bloom\nWe can turn this tunnel into room\nEvery lost line finds a tune\nMeet me down where the echoes bloom\n\n[Bridge]\nNo map, no signal, no need to run\nOur voices stack till the morning comes\n\n[Final Chorus]\nMeet me down where the echoes bloom\nWe can turn this tunnel into room\nEvery lost line finds a tune\nMeet me down where the echoes bloom"
  }
}
{
  "taskType": "audioInference",
  "taskUUID": "3431b723-fc44-450a-9ac7-2a550c58539a",
  "audioUUID": "f1e1a7b5-3aa6-41f6-811a-c5c91a832426",
  "audioURL": "https://am.runware.ai/audio/os/a01d21/ws/5/ai/f1e1a7b5-3aa6-41f6-811a-c5c91a832426.mp3",
  "seed": 39665,
  "cost": 0.00675
}