ACE-Step 1.5 music model beats Suno v5
ACE-Step 1.5 from ACE Studio and StepFun is a rare case of open-source music generation catching up with commercial AI. The model runs locally on as little as 4 GB of VRAM, produces a full track in about 2 seconds on an A100, and scores higher than Suno v5 on SongEval. Under the hood, it uses a two-stage design: an LM plans the song, and a DiT synthesizes the audio.
AI-processed from Habr AI; edited by Hamidun News
ACE-Step 1.5 aims to be the first truly strong open-source music generation model — one that doesn't just run locally but also beats Suno v5 on quality.
How can you generate music locally for free?
ACE-Step 1.5 runs on as little as 4 GB of VRAM and generates a full track in a matter of seconds.
Which music AI is better: Suno or a local model?
On SongEval, ACE-Step 1.5 outscores Suno v5, and it runs on your own computer for free, with no cloud required.
Key Questions Answered
How can you generate music locally for free?
ACE-Step 1.5 runs on as little as 4 GB of VRAM and generates a full track in a matter of seconds.
Which music AI is better: Suno or a local model?
On SongEval, ACE-Step 1.5 outscores Suno v5, and it runs on your own computer for free, with no cloud required.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.