APPLICATION

Suno: the platform that pulls the second half of music production inside one product

suno

The most complete commercial platform on the text-to-song route: a style description, your own lyrics, a hummed melody or a recorded riff all work as a starting point, a full song with vocals and instrumentation arrives in seconds, and the same product then extends it, edits sections, restyles it, extracts stems and remasters it. We list it as the top closed entry for music in the audio domain on workflow completeness rather than peak audio quality. Most generative music products cover drafting and picking a take and stop there; Suno connects the rest through stem extraction with up to 12 stems, MIDI export and Suno Studio, and Studio speaks conventional DAW semantics rather than adding another prompt box, with a multitrack timeline, take lanes, comping, manual BPM to settle tempo drift and per clip transpose and speed. Custom Models trains up to three private style variants from six or more tracks you own, and Voices, formerly Personas, generates in your own singing timbre with a verification step. The tier split matters: the free plan covers creation (generation, lyrics, Cover, crop and fade, audio upload) while stems, Add Vocals, Voices and Custom Models require Pro or Premier and Studio is Premier only on desktop web, so real cost modelling should assume Premier. Version numbers do not track quality on third party benchmarks either: on WildSongBench v5 scores 6.8721, above v6 at 6.5562 and v6 Wild at 6.4195, with v4.5 at 6.6995 and v5.5 at 6.7150, and the vendor itself asks for capability based rather than version based description. Limits: closed with no self-hosting, so unreleased melodies and lyrics must be uploaded; no stable version semantics, meaning a regeneration can sound different after a model update; commercial rights follow the tier; control granularity sits at section and style level with no editable chord track, so theory level edits require Studio multitrack re-arrangement or an open model such as YuE2 that exposes an ABC score. Graded C (vendor claim); the benchmark numbers are submitted by m-a-p and have not been recomputed by us.

CONFIDENCE: Vendor Claim; Official model card or keynote only, no independent re-test
KEY METRIC: WildSongBench SongBench 均分(v5,第三方测); Vendor Claim · 2026-09
MATURITY: Production; research → demo → product → production

Our take
We grade it C (vendor claim). The specification class facts, meaning the feature list, tier split, stem count and export formats, come from the official llms.txt dated 2026-05-29 and are verifiable but still single party statements. The WildSongBench numbers come from a third party public benchmark run by m-a-p and have not been recomputed by us, and we have run no blind listening test. The whole asset therefore carries C, and fame as the best known music product is not a reason to upgrade it.

What earns the top of the ladder is workflow completeness, not peak audio quality. Real music production runs draft, pick a take, edit sections, pull stems, re-arrange in your own DAW, master. Most generative music products cover the first two steps and stop, leaving an unusable audio file. Suno connects the rest through stem extraction with up to 12 stems, MIDI export and Suno Studio, and Studio speaks conventional DAW semantics rather than adding another prompt box: multitrack timeline, take lanes, comping, manual BPM. Manual BPM is the detail that shows someone has done real arrangement work, because tempo drift is the main reason AI material cannot enter a serious project. Custom Models, six or more owned tracks training up to three private variants, solves the other real problem, which is that style consistency cannot be maintained by rewriting a prompt every time.

Three points govern selection. First, newest is not best in this domain. On third party benchmarks v5 at 6.8721 still scores above v6 at 6.5562 and v6 Wild at 6.4195, and the vendor itself asks for capability based rather than version based description, so a version number is not a quality promise and the decision has to be made by audition. Second, control granularity sits at the section and style level, not the music theory level. There is no editable chord track and no bar level refinement, so teams needing theory level edits either re-arrange through the Studio multitrack or move to an open model such as YuE2 that exposes an ABC score as an artefact. Third, closed and not self-hostable, with commercial rights tied to tier. Unreleased melodies and lyrics have to be uploaded, which is a hard barrier where data sovereignty matters, and stems, Voices, Custom Models and Studio all sit behind paid tiers. Since those are exactly what distinguishes Suno from an ordinary generator, real cost modelling should assume the Premier tier rather than the free trial. Against Eleven Music v2.5, already in this catalogue, the Suno difference is the workstation and private style models, while the Eleven Music difference is integration with one API stack covering speech, transcription and music.

Music GenerationText-to-MusicGenerative Audio WorkstationStems

SunoOfficial site5 min read1

The problem it solves: productising everything between a one line idea and a deliverable master

Suno is the most complete commercial platform on the text-to-song route. A style description, your own lyrics, a hummed melody or a recorded riff all work as a starting point, a full song with vocals and instrumentation arrives in seconds, and then the same product extends it, edits sections, restyles it, extracts stems and remasters it. Its significance is not that it can generate songs, since both open and closed models can, but that the second half of music production is pulled inside one product. That is why this site lists it as the top closed entry for music in the audio domain.

The official material deliberately avoids version numbers (models and features ship often, so describe Suno by what it can do rather than by a version), but third party benchmarks quantify the versions: on WildSongBench, v5 scores 6.8721 on the SongBench average, v4.5 6.6995, v5.5 6.7150, v6 6.5562 and v6 Wild 6.4195. That set deserves a separate read because a newer version does not mean a higher benchmark score. On this ruler v5 remains the strongest in the family while v6 and v6 Wild sit lower. New releases usually optimise dimensions the benchmark does not measure, such as audio quality, dynamics and controllability, so newest is not automatically best in this domain and selection should be done by audition against the actual use case.

Tiering: the free tier covers creation, the paid tiers unlock production

Capability	Tier	Notes
Generation / lyrics / Cover / crop and fade / audio upload	Free	Core creation is free; Remaster can rebuild the mix of a song made with any model
Stem extraction, up to 12 stems	Pro / Premier	Vocals, drums, bass and more can be soloed, muted and downloaded as MP3, WAV, tempo-locked WAV or MIDI
Add Vocals / Add Instrumental	Pro / Premier	Layer custom vocals over an instrumental from your lyrics, or build instrumentation around a vocal idea
Voices (beta, formerly Personas)	Pro / Premier	Record or upload a short sample of your own singing and generate in your timbre, with a verification step confirming the voice is yours
Custom Models	Pro / Premier	Upload six or more original tracks you own the rights to and train up to three private variants tuned to your sound
Suno Studio (GAW)	Premier only, desktop web only	A generative audio workstation: multitrack timeline, AI stems that sit inside an existing arrangement, mic recording and transformation, loop recording with take lanes and comping, manual BPM to settle tempo drift, EQ plus per clip transpose and speed, multitrack bounce export

The last two rows are the ones to read. Custom Models and Studio turn Suno from a generator into a workstation. The first makes style controllable and private instead of re-described in every prompt; the second brings the multitrack and comping semantics of a conventional DAW while letting an AI generated stem be arranged on a timeline like any sample. Manual BPM deserves particular mention: tempo drift is the main reason generated music fails to enter a real arrangement, and locking a project to a steady tempo before export is what reconnects AI material to a DAW workflow.

Controls: three sliders and one prompt expansion

The Creative Sliders expose three continuous quantities, Weirdness (Safe to Chaos), Style Influence (Loose to Strong) and Audio Influence when working from an upload, plus a one tap prompt enhancement that expands a rough idea into a full style description. Extend adds music at the start, middle or end, with Quick Extend followed by Get Whole Song; the Song Editor allows crop and fade on any plan while replacing or adding a section inside a highlighted region belongs to advanced editing. The characteristic of this control surface is that its granularity sits at the level of sections and style, not of music theory: there is no chord track to edit.

Limits: no self-hosting, unstable version semantics, rights tied to tier

Closed, no weights, no self-hosting path. Every render passes through the Suno service, so data sovereignty cases such as unreleased melodies and lyrics have to be assessed on that basis. The self-hostable alternatives in this domain are YuE2 for music (non-commercial licence) and VoiceStudio for speech.
No stable model version semantics. The vendor asks for capability based description, and regenerating the same song at a later date can sound different after a model update. Projects needing long term timbre consistency should either commit to Voices and Custom Models or export stems and MIDI into their own DAW.
Commercial rights follow the subscription tier. Free and paid plans carry different rights over generated content, and the authoritative text is the official pricing page. This site does not restate prices or rights figures, and the vendor asks that figures not be quoted.
Benchmark scores are third party measurements of delivered audio. The WildSongBench numbers are submitted and evaluated by m-a-p and have not been recomputed by us, and they measure a delivered render rather than the human work done inside Studio.

Music Generation Text-to-Music Generative Audio Workstation Stems Suno

More in Speech & Audio

SOTA hub

Speech & AudioMediaTopC

YuE2-3B: open song generation that exposes the score as an interface

The open song generation model from m-a-p, 3B parameters, weights under CC-BY-NC-4.0, turning lyrics and a style prompt into a complete song with vocals and accompaniment at 48 kHz stereo. We list it as the open-weight top row for music in the audio domain on the strength of a combination that is close to unique among its peers: open weights plus an editable intermediate representation. One AR-NAR Mixture-of-Transformers backbone writes an ABC score (melody and chords) and semantic tokens, flow matching then produces acoustic latents and a VAE decodes them, so the score is an artefact that a person or an agent can read and edit instead of a black box whose only control is another sample. Three cot modes (full, melody, off) map onto composing, covering and direct generation, and the official agentic editing demo runs nine turns across fourteen versions from Mandarin pop to English jazz. Read the benchmark protocol carefully: on 192 WildSongBench prompts the best-of-8 SongBench average of 6.9632 sits above Suno v5 at 6.8721 and Suno v6 at 6.5562, but that is eight candidates with selection against a delivered single candidate; MuLan and AllMusicCaps, the two style-text alignment measures, still favour Suno v5, and PER at 8.44 percent trails Suno v6 Wild at 7.45 and MiniMax Music 3 at 6.27. Cost is the most concrete advantage: a 3.6 minute song in 71 seconds on an RTX 4090 24GB with a peak of 11.18 GiB, and 373 songs per hour on an H800 with vLLM at AR concurrency 32. Limits: non-commercial licence; identity preservation in covers comes almost entirely from a supplied score (CLEWS mAP collapses to 0.006 without one); the benchmark runs on the legacy VAE while the default release is the newer one; no technical report yet and no arena result. Graded C (vendor claim), not recomputed by us.

6.9632WildSongBench SongBench 均分(best-of-8)Vendor Claim · 2026-09

ResearchMultimodal Art Projects (m-a-p)Site Repo

Speech & AudioMediaTopA

VoiceStudio: the commercial voice stack moved onto local hardware and opened to agents by default

An open source, fully local voice workstation (debpalash/VoiceStudio, AGPL-3.0, Python plus Electron) positioned as the local alternative to ElevenLabs: voice cloning, voice design, video dubbing, dictation, transcription and audiobook creation, with a claimed 646 languages. Created on 2026-04-09 and at 42,831 stars and 5,005 forks on 2026-09-29. We list it as the top self-hosted row in the audio domain, not because any model behind it is stronger, but because the abstraction layer is useful. Open voice capability is scattered across a dozen repositories with their own weight formats, device requirements, cloning interfaces and licences, so assembling a pipeline of clone a voice, transcribe, align, dub and produce an audiobook spends most of its effort on glue. VoiceStudio collects 17 TTS and 11 ASR engines into one registry and reports per engine which devices it runs on (CUDA, MPS, CPU), whether it can clone, and its max_ref_seconds and ref_strategy, on top of a workflow that actually runs. The agent surface matters most for this site: a local API and an MCP server ship as first class product features, the README includes an installation prompt ready to paste into Claude Code, Codex or Cursor, and npx skills add installs it as a skill, which means coding agents can call dubbing and transcription as tools rather than needing a human at the interface. The reference audio policy is equally rigorous, with a 75 second ceiling while how much of a longer clip reaches the model depends on the engine and is reported per engine, an automatic or saved transcript ignored beyond the limit, and a hard [clone_ref_too_long] error when sending more than 20 seconds plus transcript text to the default OmniVoice. One measurement point needs stating plainly: the official docs/benchmarks.md contains a real harness with per stage profiling, refusal to start when memory is insufficient, and a rule that only RTF warm and peak memory on named hardware and versions are accepted with no estimated numbers allowed, but the results column reads No verified rows yet. We therefore attach no performance figure at all and the card carries only the star count we verified ourselves through the GitHub API, which is what earns the A grade. That A asserts the number is real and nothing about being better than ElevenLabs. Four boundaries: AGPL-3.0 is strong copyleft with a network clause, so embedding it in a service offered to outsiders obliges you to release your own server side source, and closed commercial use means either process isolation calling only the local API under your own legal review or assembling the pipeline from Apache-2.0 pieces such as CosyVoice 3 or Kokoro; per-model licences are not settled by the wrapper, since the Breeze-TTS-2 weights behind audio.cpp are research and non-commercial, Supertonic-3 and PocketTTS need extra permissions and GPT-SoVITS requires its own server, so check every model before commercial use; the 646 language figure is the repository own claim and real coverage depends on the engine chosen, from 25 European languages on Parakeet through 50 plus on FunASR to the wider Whisper family, unverified by us language by language; and desktop is Electron only now, with 0.5.3 the last Tauri release, so existing users must migrate. It sits alongside Eleven v4 on the ladder rather than replacing it, one representing the quality ceiling and convenience of a closed commercial stack and the other the data sovereignty and engine substitutability of local self-hosting.

42,831GitHub Stars(2026-09-29,本站经 GitHub API 核)Confirmed · 2026-09

Productdebpalash (open source)Site Repo

Speech & AudioMediaTopC

Eleven v4: the generation that pushes cloning fidelity far enough to reproduce bad training data

The new generation speech synthesis model from ElevenLabs, shipping in two variants: Eleven v4 for highest quality and Eleven v4 Turbo for real time use at a median inference latency of about 100 ms, with an official footnote stating that application and network latency are excluded. The vendor positions it above v3 on output quality, voice accuracy, consistency, emotion, delivery, audio tags and language coverage, and recommends migrating. We list it as the top closed source voice row in the audio domain because this generation changed the optimisation target from sounding better to sounding more like the source: timbre, cadence, delivery and other mannerisms are reproduced together, and the vendor concedes that accuracy and personal preference are not always the same thing and that you may still prefer how a voice sounded in v3. The engineering consequence is hard, since the burden of data cleaning moves back to the user. The showcase samples are explicitly labelled raw, with no EQ, compression, normalisation, de-essing or plosive removal, and the vendor says that if you hear clipping or plosives they are very likely in the training data and the model simply captured them accurately. What is deliverable comes down to three numbers: about 100 ms on Turbo; 10,000 characters per request, roughly ten minutes of audio, against 5,000 on v3 so long form text carries twice as many segment seams; and a coverage claim of 90 plus languages against a FAQ that enumerates 87, a gap we keep visible rather than smoothing away. Four boundaries matter. Cross language accent handling is a default behaviour change and not a toggle, so generating in a language that differs from the reference produces fluent native sounding speech in the target language, and the vendor calls a switchable version a research project with no timeline; any persona that depends on a native accent carried into a second language must be tested first. The control surface narrowed to Stability and Similarity, the Style and Speed sliders are gone, SSML is unsupported, and fine grained delivery moves to audio tags which the vendor admits are not perfect yet. Continuous model evolution is stated in writing, with training continuing after launch and behaviour possibly shifting over time, so teams treating a voice as a brand asset need periodic re-testing. Voice Design voices may also be less performative on v4. It is closed and not self-hostable, data must pass through ElevenLabs, and cloning compliance sits with the user; the self-hostable counterparts are CosyVoice 3, VibeVoice and Kokoro for speech and YuE2 for music under a non-commercial weight licence. This entry and the Eleven v3 already catalogued here are two generations of the same commercial stack rather than a replacement. Graded C (vendor claim): no verifiable third party benchmark, and no blind listening test by us.

~100msv4 Turbo 中位推理延迟(模型侧,不含应用与网络)Vendor Claim · 2026-09

ProductionElevenLabsSite

Speech & AudioMediaTopC

Eleven v3: the commercial audio stack that covers speaking, listening and singing

ElevenLabs' flagship speech synthesis model, officially its most emotive and expressive: 70+ languages, a 5,000-character request cap, dramatic delivery and natural multi-speaker dialogue, with a separate Text to Dialogue API making multi-character dialogue an endpoint, not a prompt trick. We list it as a top row in audio for the stack behind it: the only commercial one covering speaking, listening and singing, with four TTS tiers, three ASR tiers including a Medical build and two music models. The TTS tiers are mutually exclusive: v3 is most expressive but non-realtime at 5,000 characters, Flash v2.5 is lowest latency (about 75ms, 50% cheaper per character) and Multilingual v2 is most stable for long text at 29 languages. On ASR, Scribe v2 covers 90+ languages with diarisation for 32 speakers. Latency figures are model-side only, excluding network and upstream LLM time, not SLAs. Limits: you split long text and own continuity across the seams; closed and not self-hostable, a hard barrier where data sovereignty matters. No speech board among the ones we sync, so listed alongside Gemini 3.8 Flash TTS without ranking. Graded C (vendor-stated): no WER recomputation or blind listening test by us.

70+语言覆盖Vendor Claim · 2026-09

ProductElevenLabsSite

[●●●]LOADING

Suno: the platform that pulls the second half of music production inside one product

suno

CONFIDENCE

Vendor Claim

Official model card or keynote only, no independent re-test

6.8721

KEY METRIC

WildSongBench SongBench 均分(v5,第三方测)

Vendor Claim · 2026-09

MATURITY

Production

research → demo → product → production

Our take

We grade it C (vendor claim). The specification class facts, meaning the feature list, tier split, stem count and export formats, come from the official llms.txt dated 2026-05-29 and are verifiable but still single party statements. The WildSongBench numbers come from a third party public benchmark run by m-a-p and have not been recomputed by us, and we have run no blind listening test. The whole asset therefore carries C, and fame as the best known music product is not a reason to upgrade it.

What earns the top of the ladder is workflow completeness, not peak audio quality. Real music production runs draft, pick a take, edit sections, pull stems, re-arrange in your own DAW, master. Most generative music products cover the first two steps and stop, leaving an unusable audio file. Suno connects the rest through stem extraction with up to 12 stems, MIDI export and Suno Studio, and Studio speaks conventional DAW semantics rather than adding another prompt box: multitrack timeline, take lanes, comping, manual BPM. Manual BPM is the detail that shows someone has done real arrangement work, because tempo drift is the main reason AI material cannot enter a serious project. Custom Models, six or more owned tracks training up to three private variants, solves the other real problem, which is that style consistency cannot be maintained by rewriting a prompt every time.

Three points govern selection. First, newest is not best in this domain. On third party benchmarks v5 at 6.8721 still scores above v6 at 6.5562 and v6 Wild at 6.4195, and the vendor itself asks for capability based rather than version based description, so a version number is not a quality promise and the decision has to be made by audition. Second, control granularity sits at the section and style level, not the music theory level. There is no editable chord track and no bar level refinement, so teams needing theory level edits either re-arrange through the Studio multitrack or move to an open model such as YuE2 that exposes an ABC score as an artefact. Third, closed and not self-hostable, with commercial rights tied to tier. Unreleased melodies and lyrics have to be uploaded, which is a hard barrier where data sovereignty matters, and stems, Voices, Custom Models and Studio all sit behind paid tiers. Since those are exactly what distinguishes Suno from an ordinary generator, real cost modelling should assume the Premier tier rather than the free trial. Against Eleven Music v2.5, already in this catalogue, the Suno difference is the workstation and private style models, while the Eleven Music difference is integration with one API stack covering speech, transcription and music.

Music GenerationText-to-MusicGenerative Audio WorkstationStems

The problem it solves: productising everything between a one line idea and a deliverable master

Tiering: the free tier covers creation, the paid tiers unlock production

Capability

Tier

Notes

Generation / lyrics / Cover / crop and fade / audio upload

Free

Core creation is free; Remaster can rebuild the mix of a song made with any model

Stem extraction, up to 12 stems

Pro / Premier

Vocals, drums, bass and more can be soloed, muted and downloaded as MP3, WAV, tempo-locked WAV or MIDI

Add Vocals / Add Instrumental

Pro / Premier

Layer custom vocals over an instrumental from your lyrics, or build instrumentation around a vocal idea

Voices (beta, formerly Personas)

Pro / Premier

Record or upload a short sample of your own singing and generate in your timbre, with a verification step confirming the voice is yours

Custom Models

Pro / Premier

Upload six or more original tracks you own the rights to and train up to three private variants tuned to your sound

Suno Studio (GAW)

Premier only, desktop web only

A generative audio workstation: multitrack timeline, AI stems that sit inside an existing arrangement, mic recording and transformation, loop recording with take lanes and comping, manual BPM to settle tempo drift, EQ plus per clip transpose and speed, multitrack bounce export

Controls: three sliders and one prompt expansion

Limits: no self-hosting, unstable version semantics, rights tied to tier

Closed, no weights, no self-hosting path. Every render passes through the Suno service, so data sovereignty cases such as unreleased melodies and lyrics have to be assessed on that basis. The self-hostable alternatives in this domain are YuE2 for music (non-commercial licence) and VoiceStudio for speech.

No stable model version semantics. The vendor asks for capability based description, and regenerating the same song at a later date can sound different after a model update. Projects needing long term timbre consistency should either commit to Voices and Custom Models or export stems and MIDI into their own DAW.

Commercial rights follow the subscription tier. Free and paid plans carry different rights over generated content, and the authoritative text is the official pricing page. This site does not restate prices or rights figures, and the vendor asks that figures not be quoted.

Benchmark scores are third party measurements of delivered audio. The WildSongBench numbers are submitted and evaluated by m-a-p and have not been recomputed by us, and they measure a delivered render rather than the human work done inside Studio.