{"slug": "ace-step-1-5", "name": "ACE-Step 1.5", "category": "ai-models", "category_name": "AI models", "description": "", "recommenders": 3, "commenters": 4, "warners": 1, "score": 1.733, "rising": -0.98, "first": "2026-02-09", "latest": "2026-04-07", "why": "the results were actually good with it being opensource and for how cheap the model is to run.", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 3, "neutral": 0, "compare": 0, "warn": 1}, "spark": [0, 0, 0, 2, 0, 1, 0, 0, 0, 0, 0, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "the results were actually good with it being opensource and for how cheap the model is to run.", "kind": "", "icon": true, "domain": null, "github_repo": "github.com/ace-step/ace-step-1.5", "official_url": "https://github.com/ace-step/ACE-Step-1.5", "pricing": "", "aliases": ["ACE-Step 1.5", "ACE-Step", "Ace-Step-1.5"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 762823168, "item_id": 47676448, "url": "https://news.ycombinator.com/item?id=47676448", "by": "mesmertech", "date": "2026-04-07", "version": "1.5", "stance": "recommend", "quote": "the results were actually good with it being opensource and for how cheap the model is to run.", "excerpt": "Was playing around with it earlier and the results were actually good with it being opensource and for how cheap the model is to run. The tool page itself is smth I vibecoded in like an hour. Feel free to give it a try", "context": "", "use_case": "music generation", "switched_from": "", "story": "Ace-Step-1.5 playground- SOTA open source music model"}, {"id": 753170901, "item_id": 47073181, "url": "https://news.ycombinator.com/item?id=47073181", "by": "tarruda", "date": "2026-02-19", "version": null, "stance": "recommend", "quote": "ACEStep (very good music generation model)", "excerpt": "This is probably one of the most underrated LLMs releases in the past few months. In my local testing with a 4-bit quant (https://huggingface.co/ubergarm/Step-3.5-Flash-GGUF/tree/main/IQ4_XS), it surpasses every other LLM I was able to run locally, including Minimax 2.5 and GLM-4.7, though I was only able to run GLM with a 2-bit quant. Some highlights: - Very context efficient: SWA by default, on a 128G mac I can run", "context": "", "use_case": "", "switched_from": "", "story": "Step 3.5 Flash: Fast Enough to Think. Reliable Enough to Act"}, {"id": 752817696, "item_id": 47051106, "url": "https://news.ycombinator.com/item?id=47051106", "by": "falloon", "date": "2026-02-17", "version": "1.5", "stance": "recommend", "quote": "Ace step 1.5 is great, only 1.5b params so very easy to run locally.", "excerpt": "Ace step 1.5 is great, only 1.5b params so very easy to run locally. https://github.com/ace-step/ACE-Step-1.5", "context": "", "use_case": "", "switched_from": "", "story": "Claude Sonnet 4.6"}, {"id": 751175744, "item_id": 46948484, "url": "https://news.ycombinator.com/item?id=46948484", "by": "DanielWen", "date": "2026-02-09", "version": "1.5", "stance": "recommend", "quote": "experimenting with ACE-Step 1.5 lately", "excerpt": "I\u2019ve been experimenting with ACE-Step 1.5 lately and wanted to share a short summary of what actually helped me get more controllable and musical results, based on the official tutorial + hands-on testing. The biggest realization: ACE-Step works best when you treat prompts as [structured inputs], not a single sentence (same as other LLMs) 1. Separate \u201cTags\u201d from \u201cLyrics\u201d Instead of writing one long prompt, think in t", "context": "", "use_case": "Music generation", "switched_from": "", "story": "Ace-Step 1.5 prompt tips: how I get more controllable music output"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 751246656, "item_id": 46952916, "url": "https://news.ycombinator.com/item?id=46952916", "by": "Rochus", "date": "2026-02-09", "version": null, "stance": "warn", "quote": "the musicality of this system is much lower than what we have commercially", "excerpt": "I didn't try what you suggested, but even without better control of the music with more structured prompts, I think the musicality of this system is much lower than what we have commercially. I made the following two songs with the system: - http://rochus-keller.ch/Diverses/Ace-Step-v1.5_demo1.mp3 - http://rochus-keller.ch/Diverses/Ace-Step-v1.5_demo2.mp3 with the same prompt as e.g. this song https://rochus-keller.c", "context": "", "use_case": "Music generation", "switched_from": "", "story": "Ace-Step 1.5 prompt tips: how I get more controllable music output"}]}], "who": [], "tags": [], "use_cases": ["music generation"], "compared": [{"slug": "suno", "name": "Suno", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [{"id": 749982288, "item_id": 46873893, "url": "https://news.ycombinator.com/item?id=46873893", "by": "DanielWen", "date": "2026-02-03", "version": "1.5", "stance": "recommend", "quote": "We released ACE-Step 1.5, an open-sourced AI music model.", "excerpt": "Hi HN! We released ACE-Step 1.5, an open-sourced AI music model. Key traits of ACE-Step 1.5: Quality: beats Suno on common eval scores Speed: full song under 2s on A100 Local: ~4GB VRAM, under 10s on RTX 3090 LoRA: train your own style with a few songs License: MIT, free for commercial use Data: fully authorized plus synthetic GitHub: https://github.com/ace-step/ACE-Step-1.5 Weights/Training code/LoRA code/Paper are ", "context": "", "use_case": "", "switched_from": "", "story": "SHOW HN:AceStep1.5an on-device music model that beats Suno on common eval metric"}], "threads": [{"url": "https://news.ycombinator.com/item?id=46948483", "title": "Ace-Step 1.5 prompt tips: how I get more controllable music output", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=47676447", "title": "Ace-Step-1.5 playground- SOTA open source music model", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47069179", "title": "Step 3.5 Flash: Fast Enough to Think. Reliable Enough to Act", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47050447", "title": "Claude Sonnet 4.6", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=46873892", "title": "SHOW HN:AceStep1.5an on-device music model that beats Suno on common eval metric", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {"warn": 1, "recommend": 2}, "2026-03": {}, "2026-04": {"recommend": 1}, "2026-05": {}, "2026-06": {}, "2026-07": {}, "2026-08": {}, "2026-09": {}, "2026-10": {}}, "versions": [], "version_unnamed": 2, "family_ref": null, "runs_on": null, "lines": [], "quotes": [{"id": 762823168, "item_id": 47676448, "url": "https://news.ycombinator.com/item?id=47676448", "by": "mesmertech", "date": "2026-04-07", "version": "1.5", "stance": "recommend", "quote": "the results were actually good with it being opensource and for how cheap the model is to run.", "excerpt": "Was playing around with it earlier and the results were actually good with it being opensource and for how cheap the model is to run. The tool page itself is smth I vibecoded in like an hour. Feel free to give it a try", "context": "", "use_case": "music generation", "switched_from": "", "story": "Ace-Step-1.5 playground- SOTA open source music model"}, {"id": 753170901, "item_id": 47073181, "url": "https://news.ycombinator.com/item?id=47073181", "by": "tarruda", "date": "2026-02-19", "version": null, "stance": "recommend", "quote": "ACEStep (very good music generation model)", "excerpt": "This is probably one of the most underrated LLMs releases in the past few months. In my local testing with a 4-bit quant (https://huggingface.co/ubergarm/Step-3.5-Flash-GGUF/tree/main/IQ4_XS), it surpasses every other LLM I was able to run locally, including Minimax 2.5 and GLM-4.7, though I was only able to run GLM with a 2-bit quant. Some highlights: - Very context efficient: SWA by default, on a 128G mac I can run", "context": "", "use_case": "", "switched_from": "", "story": "Step 3.5 Flash: Fast Enough to Think. Reliable Enough to Act"}, {"id": 752817696, "item_id": 47051106, "url": "https://news.ycombinator.com/item?id=47051106", "by": "falloon", "date": "2026-02-17", "version": "1.5", "stance": "recommend", "quote": "Ace step 1.5 is great, only 1.5b params so very easy to run locally.", "excerpt": "Ace step 1.5 is great, only 1.5b params so very easy to run locally. https://github.com/ace-step/ACE-Step-1.5", "context": "", "use_case": "", "switched_from": "", "story": "Claude Sonnet 4.6"}, {"id": 751246656, "item_id": 46952916, "url": "https://news.ycombinator.com/item?id=46952916", "by": "Rochus", "date": "2026-02-09", "version": null, "stance": "warn", "quote": "the musicality of this system is much lower than what we have commercially", "excerpt": "I didn't try what you suggested, but even without better control of the music with more structured prompts, I think the musicality of this system is much lower than what we have commercially. I made the following two songs with the system: - http://rochus-keller.ch/Diverses/Ace-Step-v1.5_demo1.mp3 - http://rochus-keller.ch/Diverses/Ace-Step-v1.5_demo2.mp3 with the same prompt as e.g. this song https://rochus-keller.c", "context": "", "use_case": "Music generation", "switched_from": "", "story": "Ace-Step 1.5 prompt tips: how I get more controllable music output"}, {"id": 751175744, "item_id": 46948484, "url": "https://news.ycombinator.com/item?id=46948484", "by": "DanielWen", "date": "2026-02-09", "version": "1.5", "stance": "recommend", "quote": "experimenting with ACE-Step 1.5 lately", "excerpt": "I\u2019ve been experimenting with ACE-Step 1.5 lately and wanted to share a short summary of what actually helped me get more controllable and musical results, based on the official tutorial + hands-on testing. The biggest realization: ACE-Step works best when you treat prompts as [structured inputs], not a single sentence (same as other LLMs) 1. Separate \u201cTags\u201d from \u201cLyrics\u201d Instead of writing one long prompt, think in t", "context": "", "use_case": "Music generation", "switched_from": "", "story": "Ace-Step 1.5 prompt tips: how I get more controllable music output"}]}