ElevenLabs Music
Last reviewed
Jun 5, 2026
Sources
24 citations
Review status
Source-backed
Revision
v4 ยท 5,876 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 5, 2026
Sources
24 citations
Review status
Source-backed
Revision
v4 ยท 5,876 words
Add missing citations, update stale details, or suggest a clearer explanation.
ElevenLabs Music (also marketed as Eleven Music) is an AI music generation product developed by ElevenLabs, the voice AI company founded in 2022 by Piotr Dabkowski and Mati Staniszewski. It allows users to generate studio-grade music from natural language prompts, covering vocals and instrumentals across many genres and languages. The product launched in August 2025 alongside licensing agreements with independent music rights holders, making it one of the first commercial AI music generators to secure such deals before going public.[1][3] A consumer-facing iOS app, ElevenMusic, followed in April 2026 and expanded the platform into streaming and social music discovery.[6][9] In late May 2026 the company released its second-generation model, Music v2, which added section-level inpainting, mid-song genre transitions, and embedded sound effects, and powers ElevenMusic, the ElevenLabs API, and ElevenCreative.[19][20]
ElevenLabs built its reputation on text-to-speech. When the company launched in January 2023, its voice synthesis models produced unusually natural-sounding output, capable of conveying emotion, laughter, and cadence in ways that earlier systems from Apple or Amazon could not. Within six months the platform had crossed one million registered users. Funding came quickly: a $19 million Series A in June 2023, an $80 million Series B in January 2024 led by Andreessen Horowitz and Sequoia Capital, and a $500 million Series C in February 2026 that pushed the company's valuation to $11 billion.
As the voice synthesis market matured, ElevenLabs moved into adjacent audio categories. It shipped an AI dubbing studio in October 2023, a mobile reader app in June 2024, and a sound effects generator shortly after. Music was a natural next step. The company first shared internal samples from a text-to-music model in May 2024, and limited public access opened in July 2025 before the full commercial launch in August 2025.[1]
The timing was deliberate. By mid-2025, the legal landscape for AI music had grown contentious. The Recording Industry Association of America (RIAA) had sued Suno and Udio in June 2024 for alleged mass infringement of copyrighted recordings used in training. Both companies were in settlement negotiations with major labels through late 2025. ElevenLabs chose a different path: it entered the market with licensing deals already in place rather than waiting for lawsuits to force the issue.[1][3]
The music product is also part of a strategic hedge for ElevenLabs. Company leadership has acknowledged that voice synthesis faces eventual commoditization as models from OpenAI, Google, and Meta continue to improve. Expanding into music, sound effects, and video creates a broader audio platform that is harder to displace than a single-product voice tool. Mati Staniszewski described music as an area where ElevenLabs sees "tools that use AI to create music and other mediums as a way to grow and protect itself from the eventual commoditization of AI audio models."[17]
| Date | Event |
|---|---|
| May 2024 | Internal samples from text-to-music model shared publicly |
| July 2024 | Limited public beta opens; basic music generation available to users |
| July 23, 2025 | SourceAudio named preferred music licensing partner [5] |
| August 5, 2025 | Full Eleven Music launch with commercial licensing; Merlin and Kobalt deals announced [1][2] |
| January 21, 2026 | "The Eleven Album" released on Spotify featuring Liza Minnelli, Art Garfunkel, and others [11] |
| March 19, 2026 | Music Marketplace launched; creators can monetize AI-generated tracks [10] |
| April 1, 2026 | ElevenMusic iOS app releases on Apple App Store [6][14] |
| April 29, 2026 | ElevenMusic app updated to version 1.0.8 |
| April 30, 2026 | ElevenMusic rebranded and expanded as a streaming and fan engagement platform [8][9] |
| May 26, 2026 | Music v2 released; pricing cut by up to 50% on the API and up to 40% on ElevenCreative self-serve [19] |
The copyright strategy behind Eleven Music is the feature that most distinguishes it from competitors at launch. ElevenLabs built the initial model using licensed production music rather than scraped recordings, and it structured deals with three organizations before the public launch.[2][3]
In July 2025, ElevenLabs named SourceAudio its preferred music licensing partner.[5] SourceAudio is a sync licensing platform used by more than 4,000 music catalogs, including independent labels and production music libraries. At the time of the partnership, approximately 14 million songs on the SourceAudio platform had opted into its AI dataset licensing program.[5] The agreement gave ElevenLabs access to that catalog subject to conditions including no direct song replication and no generation of content intended to substitute for the licensed material.[5]
Merlin is the digital licensing organization for independent labels and distributors. It represents more than 30,000 member companies whose catalogs include artists such as Adele, Nirvana, Mitski, Carly Rae Jepsen, and Phoebe Bridgers.[2] Under the ElevenLabs deal, Merlin members can voluntarily opt their master recordings into AI training in exchange for a share of a royalty pool.[2] The agreement includes specific safeguards: Eleven Music users cannot enter artist names, songwriter names, song titles, or album titles as generation prompts, and the system cannot produce outputs designed to mimic identifiable artists.[2] Merlin CEO Anna Camu described the framework as creating "responsible guardrails" for the AI music space.[2]
Kobalt Music Group is one of the largest independent music publishers, representing songwriters behind artists including Beck, Bon Iver, and Childish Gambino.[2] Kobalt's deal with ElevenLabs covers publishing rights (mechanical and synchronization) on a two-year global term.[4] Kobalt CEO Laurent Hubert stated that "artists have to voluntarily opt in for their music to be licensed for AI use" and that "clients benefit directly" through revenue sharing.[4]
The royalty structure under the Kobalt arrangement is a 50/50 split between publishing (composition rights) and recorded music (master rights), a parity model considered favorable to songwriters who often receive a smaller share in traditional licensing.[2] Participating creators receive a pro-rata share of a royalty pool based on how many of their works were used in training, weighted by digital popularity metrics on streaming platforms.[2] Eligibility is currently limited to creators whose master recordings and publishing copyrights are both 100% controlled by Merlin members and Kobalt, respectively.[2]
A separate tier, Eleven Music Pro, was announced in August 2025 and described as a subsequent model to be trained specifically on the opted-in Merlin and Kobalt catalog.[2] The timeline for releasing Eleven Music Pro had not been publicly confirmed as of May 2026.
In May 2026, alongside the launch of Music v2, ElevenLabs added a licensing collaboration with Believe, the digital music company that distributes and develops independent artists and labels worldwide.[19][21] ElevenLabs cited the Believe agreement as part of its position that Music v2 is trained only on licensed data and cleared for commercial use.[19] The deal extended the pool of opted-in catalog beyond the original Merlin, Kobalt, and SourceAudio partners.[21]
Alongside the licensing deals, ElevenLabs implemented prompt-level restrictions to prevent outputs that could infringe existing works. Users are prohibited from submitting prompts containing artist or songwriter names, song or album titles, publisher or label names, or substantial portions of song lyrics.[2] The terms of service also prohibit using generated music in arms, tobacco, pharmaceutical, adult entertainment, religious, and political contexts without prior written approval from ElevenLabs.[13] Standard commercial use (film, television, podcasts, social media, advertising, gaming) is covered under paid subscription tiers.[13]
Eleven Music accepts natural language descriptions of desired music and generates full audio output. Users can specify genre, mood, tempo, instrumentation, and vocal style in plain text without needing musical vocabulary, though the system also understands standard music terminology.[13] The model supports custom lyrics supplied by the user or can generate its own lyrics from the prompt context.[13]
Generated tracks can include vocals or be purely instrumental. Vocal output is available in English, Spanish, German, and Japanese.[13] Tracks range in duration from 3 seconds to 5 minutes, with an Auto setting that determines length dynamically based on the prompt.[13]
Users can construct songs piece by piece rather than generating a complete track in one pass. The interface allows defining distinct sections including intros, verses, choruses, breakdowns, and outros.[13] Each section can be regenerated independently, so a weak chorus can be replaced without rebuilding the whole song.[13] Lyrics can also be edited at the section level. This workflow is closer to song arrangement than to single-shot generation. With the Music v2 release in May 2026, section regeneration was upgraded to true inpainting, which leaves the audio outside the edited section unchanged at the sample level (see Music v2 below).[19][20]
Generated music is exported as MP3 at 44.1kHz, 128 to 192kbps.[13] Higher subscription tiers unlock studio-grade exports with full dynamic range. Stem export is available for up to six tracks, enabling users to take individual elements such as drums, bass, melody, and vocals into an external digital audio workstation for further production.[13] ElevenLabs has described stems as including drums, bass, melody, harmony, and vocals, giving producers granular control over the mix.[13]
Eleven Music includes a finetune system that lets creators train a custom version of the base model on their own audio.[13] A creator uploads non-copyrighted tracks they own, names the finetune, adds tags, and sets sharing permissions.[13] Training takes approximately 5 to 10 minutes.[13] The resulting finetune captures stylistic patterns across instrumentation, tempo, production style, timbre, and vocal character, and can be selected during generation to produce output consistent with that style. Finetunes can be shared within a workspace so collaborators can access the same custom model.
The platform also ships with curated finetunes covering global genres and production styles for users who do not want to train their own. Enterprise subscribers with proprietary catalogs can fine-tune on their own intellectual property without the third-party copyright screening that applies to other tiers. For standard users, all uploaded audio is automatically screened by a third-party copyright compliance system before fine-tuning begins.[13] At the time of the Music v2 launch, finetunes were available only for the v1 model, with v2 finetune support described as forthcoming.[19]
The platform supports generation of tracks up to 8 minutes in some configurations, which is longer than the typical 2 to 4 minute limits on competing platforms. This is relevant for ambient and background music that needs to loop or sustain without audible stitching artifacts. A style transfer feature allows users to upload a reference audio clip and generate new compositions that match the tonal qualities of that reference without copying specific melodies or arrangements.
A music generation API is available to paid subscribers, allowing developers to integrate Eleven Music into applications, games, and automated workflows.[13] The API is part of ElevenLabs' broader audio platform, which also exposes text-to-speech, sound effects, and voice cloning endpoints through a unified interface. All capabilities are accessible through a single API key, and credits are shared across products, so unused music credits can be applied to voice generation and vice versa.
On May 26, 2026, ElevenLabs released Music v2, the second generation of its music model, roughly ten months after the original August 2025 launch.[19][20] The company described it as delivering better vocals, instrumentation, and arrangement across every genre, and positioned it as the new default model in the ElevenMusic interface while keeping Music v1 available during the transition.[19] Music v2 powers three products: ElevenMusic, the consumer app for listening, remixing, and creating tracks; the ElevenLabs API, for embedding music generation in other products; and ElevenCreative, which provides downloadable music for advertising, branded content, and video.[19]
The headline addition is inpainting. A user can select a specific section of a generated track and regenerate only that part, leaving the rest of the song unchanged at the sample level.[19][23] ElevenLabs gave the example of rewriting a song's bridge without altering the verses, chorus, or outro, comparing the workflow to non-destructive editing of individual regions in a digital audio workstation such as Pro Tools or Logic.[19] This differs from the earlier section-regeneration workflow, where regenerating one section could shift the surrounding audio.
Music v2 can also transition between distinct genres inside a single track while holding musical coherence.[20][21] ElevenLabs cited moving from opera to heavy metal and back as an example, and other demonstrations included acoustic folk shifting into dubstep and jazz piano into an industrial breakbeat.[20][21] The model handles fast rap and dense lyrical delivery without losing structure, and it can embed non-musical sound effects directly inside a track rather than requiring them to be layered in afterward.[19][23]
Long-form, section-by-section composition lets users build a complete song one part at a time, defining intros, verses, choruses, and other sections while the model maintains continuity across them.[19] ElevenLabs also reported improved multilingual performance in v2, covering vocals, lyrics, and arrangement across a growing set of languages.[19] The supported languages include English, Spanish, German, and Japanese among others.[13] According to the company's documentation, Music v2 improves on Music v1 in prompt adherence, composition, prompt understanding, and multilingual output.[22]
ElevenLabs stated that Music v2 is trained only on licensed data and cleared for commercial use, so generated tracks carry no sync fees, clearance delays, or restrictions on deployment.[19][23] The release coincided with a new licensing collaboration with Believe, which joined the existing arrangements with Merlin, Kobalt, and SourceAudio.[19][21] The commercial-use position is consistent with the company's original licensing-first strategy and continues to distinguish ElevenLabs from rivals whose training data drew legal challenges.[19]
Alongside the Music v2 launch, ElevenLabs cut music generation prices for both v1 and v2.[19] API pricing dropped by up to 50%, and ElevenCreative self-serve pricing dropped by up to 40%.[19] Reported API rates were around $0.15 per minute of generated audio after the cut.[20] The maximum single-track length remained 5 minutes, and audio output remained MP3 at 44.1kHz, 128 to 192kbps, with WAV also available.[13]
In January 2026, ElevenLabs released "The Eleven Album," a two-volume project described as the first major AI-assisted album produced with the full voluntary participation of the featured artists.[11][12] The album was available to stream on both elevenlabs.io and Spotify.[11]
Artists contributing tracks included Liza Minnelli, Art Garfunkel, Patrick Patrikios (a producer who has worked with Sia, Nicki Minaj, and Britney Spears), Willonius, IAMSU!, Demitri Leiros, Emily Falvey, Sunsetto, Kondzilla, Chris Lyons, and Michael Feinstein.[11][12] The album spans rap, pop, R&B, and EDM.[11]
Minnelli's contribution, titled "Kids Wait Till You Hear This," is an EDM track. Her representative confirmed that the track uses her actual recorded voice rather than an AI reproduction.[11] Minnelli said of the project: "What interested me here was the idea of using my voice and new tools in service of expression, not instead of it. This project respects the artist's voice, the artist's choices, and the artist's ownership."[11]
Garfunkel contributed a spoken-word piece called "Authorship," backed by piano and ambient sound. He noted: "Music has always evolved alongside technology. What impressed me about this experience was the respect for musicianship. The human remains at the center."[11]
All streaming revenue from the album flows back to the participating artists under ElevenLabs' creator-control framework.[11][12]
On March 19, 2026, ElevenLabs launched the Music Marketplace, a platform feature allowing users to publish AI-generated tracks and earn revenue when other users license or use those tracks commercially.[10] The system mirrors the Voice Marketplace ElevenLabs had operated for voice creators, through which it had paid more than $11 million to creators by the time of the music launch.[10]
Paid subscribers can browse published tracks and license them for use in advertisements, games, videos, and other commercial applications.[10] When a paid subscriber uses a creator's track, revenue flows to the creator.[10] ElevenLabs co-founder Mati Staniszewski said at the marketplace launch that "our community has already created fourteen million songs with Eleven Music" since the August 2025 debut.[10]
The Marketplace also allows tracks to be remixed. A user can take a published song and modify it through a new text prompt to change genre, tempo, or instrumentation, with the result counted as a new generation.[10] Early notable participants included Patrick Jordan-Patrikios, a producer whose credits include work for Sia, Nicki Minaj, and Britney Spears.[10]
On April 1, 2026, ElevenLabs released ElevenMusic on the Apple App Store.[6][14] The app had been listed for several weeks before going live.[6] It brought the Eleven Music generation capabilities to mobile for the first time, alongside a streaming and music discovery layer not present in the web platform.[6][7]
Users generate songs using natural language prompts within the app.[6] Controls include song length, whether the track has lyrics, and writing style.[6] Each remix counts toward the user's daily generation limit. The free tier allows up to seven song creations per day.[6] A Pro subscription at $9.99 per month (or $95.90 per year) unlocks up to 500 monthly track creations, 500 GB of storage, and full access to all styles and moods.[14] Songs created in the app can be shared with customizable visual displays built around the track.
ElevenMusic functions as a streaming service, not just a generation tool.[8][9] The app includes live stations, pre-created albums, and daily mixes organized by mood: Focus, Energy, Relax, Late Night, Cosmic, and Chill.[9] Users can browse top charts, trending tracks, and new releases.[9] Songs created by other users can be discovered within the app and remixed via text prompts.[9]
At the time of the app's consumer relaunch on April 30, 2026, the platform had approximately 4,000 human artists, primarily emerging acts.[8][9] Derek Cournoyer, ElevenLabs' music strategy lead, described the platform's goal as giving "fans a more active way to experience the music they love, and create a new gateway into DJing."[9] The platform was positioned not as a listening app but as a "fan engagement layer" that transforms passive listeners into active participants in the creative process.[8][9]
The app requires iOS 18.0 or later.[14] It runs on iPhone, iPad, Mac devices with Apple M1 chips or newer, and Apple Vision.[14] It is available in English, German, Polish, Portuguese, and Spanish.[14] The app size is 34.4 MB.[14] The age rating is 13+.[14]
ElevenLabs Music is accessible within the broader ElevenLabs subscription structure, which uses a unified credit system across voice, sound effects, and music. Music generation is billed per second of audio generated rather than per track, which means longer tracks consume more credits than short ones. Unused credits within a billing period do not roll over on standard plans but can be applied to other ElevenLabs products (voice generation, sound effects) within the same billing cycle.
| Plan | Monthly price | Music generation included | Commercial license |
|---|---|---|---|
| Free | $0 | Approximately 11 minutes | No (attribution required) |
| Starter | $5 | Approximately 30 minutes | Yes |
| Creator | $22 | Approximately 62 minutes | Yes |
| Pro | $99 | Approximately 304 minutes | Yes |
| Scale | $330 | Higher volume | Yes |
| Enterprise | Custom | Custom | Yes |
The ElevenMusic iOS app has its own separate pricing tier: free accounts generate up to seven tracks per day, and the $9.99 monthly Pro tier offers up to 500 monthly creations.[14] Annual billing at $95.90 saves approximately 20% over monthly pricing.[14]
Additional music can be generated at overage rates, charged at approximately $0.50 per minute of generated audio on paid plans. For the free tier, generation beyond the daily limit is not available without upgrading. Annual billing across all ElevenLabs plans saves approximately 17% compared to monthly billing, equivalent to roughly two free months per year.
In May 2026, ElevenLabs reduced music generation pricing for both the v1 and v2 models.[19] The cut reached up to 50% for API customers and up to 40% for ElevenCreative self-serve customers, with reported API rates of around $0.15 per minute of generated audio after the change.[19][20]
Eleven Music is part of ElevenCreative, ElevenLabs' broader creative platform that consolidates voice synthesis, music generation, sound effects, dubbing, and audio-visual editing under a single workspace.[19] Music outputs connect directly with the voice and sound effects tools within a built-in audio editor, allowing a creator to generate a voice narration, a background music track, and ambient sound effects, then mix all three in the same environment.
This integration is a practical differentiator for users who already pay for ElevenLabs voice tools. A podcast producer, for example, can generate intro music, a narrated script using a cloned voice, and transition sounds in one workflow without exporting to a third-party editor. For video producers, a unified API call chain can produce complete audio sequences covering narration, score, and effects.
The platform supports multi-seat workspaces with shared credit pools and role-based access, which allows production teams to manage audio generation collaboratively without each member holding a separate subscription.
The three main AI music generation platforms as of mid-2026 are Eleven Music, Suno (on version 4 and evolving toward v5), and Udio.[15][16] Each platform has developed distinct strengths. The May 2026 release of Music v2 narrowed several of the gaps described below, in particular by adding inpainting and improving vocals; the ratings in the table reflect independent testing conducted largely on the v1 model and earlier.[15][16]
| Feature | ElevenLabs Music | Suno | Udio |
|---|---|---|---|
| Licensed training data | Yes (Merlin, Kobalt, SourceAudio, Believe) | No (RIAA lawsuit settled 2025) | No (RIAA lawsuit settled 2025) |
| Vocal quality | Moderate (improved in v2) | Strong | Moderate |
| Instrumental quality | Strong | Good | Excellent |
| Cinematic/orchestral | Moderate | Limited | Best in class |
| Section editing | Yes (inpainting in v2) | Limited | Yes (inpainting) |
| Mid-song genre transitions | Yes (v2) | Limited | Limited |
| Embedded sound effects | Yes (v2) | No | No |
| Stem export | Yes (up to 6 stems) | No | Yes |
| Custom finetunes | Yes (v1; v2 forthcoming) | No | No |
| Mobile app | iOS (April 2026) | iOS and Android | iOS and Android |
| Generation speed | 25-30 sec (instrumental) | ~43 sec average | 5-10 minutes |
| Commercial license clarity | Highest | Lower (litigation history) | Lower (litigation history) |
| Free tier songs/day | 7 (ElevenMusic app) | 5 | 5 |
| Pro plan price | $22/mo (Creator) | $10/mo | $10/mo |
Suno leads in overall song coherence, particularly for pop and rock with vocals.[15] Its interface produces complete songs with recognizable verse-chorus-bridge structure more reliably than competitors. Independent testing has found its vocal synthesis to score highest for naturalness among the three platforms.[15]
Udio's core differentiator at the time of those tests was inpainting: the ability to regenerate a specific section of a track without affecting surrounding audio.[15][16] Combined with stem separation and manual extension controls, it offered the most fine-grained editing of the three. With Music v2 in May 2026, ElevenLabs added its own inpainting, narrowing that gap.[19][20] For cinematic and orchestral music, Udio produces depth and spatial dynamics that the other platforms do not match.[16] Its audio output fidelity, at 48kHz/32-bit, is higher than ElevenLabs' 44.1kHz standard export.[16]
ElevenLabs Music occupies a different position. Its vocal output in isolated tests on the v1 model was rated lower than Suno's by multiple independent reviewers, and the company says v2 improves vocals across genres.[15] In a six-month comparative study published in 2026, only 3 of 20 vocal track requests produced usable results, with the majority described as emotionally flat.[15] Its instrumental generation is faster and more consistent, particularly for background music, ambient scoring, and podcast beds. One reviewer noted that ElevenLabs "cranks these out in 25-30 seconds" for ambient and corporate scoring tasks.[15]
The primary advantage ElevenLabs offers is its licensing clarity. Both Suno and Udio faced RIAA lawsuits over their training data and reached settlements with major labels in late 2025.[15][16] ElevenLabs entered the market with deals already in place, which gives its commercial outputs a lower legal risk profile for creators working with advertisers, broadcast clients, or platforms with strict Content ID policies.[1] Reviewers rated ElevenLabs highest for commercial license clarity at 9.5 out of 10, compared to 6 out of 10 for both Suno and Udio.[16]
The pricing model also differs. Suno and Udio charge per song with generous monthly song caps at competitive prices ($10/month for Pro).[15][16] ElevenLabs charges per minute of generated audio, which integrates with its broader credit system and allows unused music credits to be applied to voice or sound effects. The May 2026 price cuts lowered API rates to around $0.15 per minute.[20] For users already paying for ElevenLabs voice tools, the marginal cost of adding music is lower. For users who only want music, Suno and Udio offer more songs per dollar at their entry price points.
ElevenLabs operates voice cloning as a separate product from Eleven Music. The voice cloning system, which includes both instant cloning (from a 10-second sample) and professional cloning (from a larger dataset), is not currently integrated with the music generation model. Professional Voice Clones support only spoken audio and do not transfer singing capabilities to the music model.[13]
The integration between voice and music operates at the platform level rather than the model level. A creator can generate a voiceover using a cloned voice and a music bed using Eleven Music, then combine them in the ElevenLabs project editor. This workflow is relevant for podcast production and video narration where background music and spoken audio need to match in tone and pacing.
The absence of singing capability in voice clones is an explicit limitation in ElevenLabs' documentation. The company notes that Professional Voice Clones require audio recordings consisting of spoken voice only, and that singing is not supported.[13] This separates ElevenLabs from some specialized vocal AI tools that do allow singing synthesis from cloned voice profiles.
Eleven Music targets several distinct user categories.
Content creators and video producers use it for background music in YouTube videos, social media content, and tutorials. The commercial license and low Content ID risk are relevant for monetized channels where disputed copyright claims can demonetize content or trigger takedowns.
Podcast producers generate intro and outro music, transitions, and ambient beds. Background music can be tuned at the stem level to stay behind speech frequencies. The fast instrumental generation suits high-volume workflows where individual track creation needs to scale across many episodes.
Game developers and interactive media teams use the API to generate situational music at scale. The ability to specify mood, tempo, and instrumentation through text prompts fits procedural audio pipelines where music needs to vary dynamically based on game state.
Advertising and brand work represents the highest-value commercial use case. ElevenLabs has worked with gaming, advertising, music, and media companies for custom music generation. The 50/50 royalty parity structure and explicit commercial clearance make the music usable in broadcast contexts where other AI music tools carry residual legal uncertainty.[2] Short-form audio branding, such as 3-second audio logos with extended looping variations, can be produced efficiently through the platform.
Artists and producers use the finetune system to capture a sonic identity and generate variations in that style, or to produce reference tracks quickly before committing to full manual production. Six-stem export allows work started in Eleven Music to be continued in a digital audio workstation.[13] The inpainting added in Music v2 lets them revise a single section of a near-final track without re-rolling the entire song.[19]
Film and video scoring workflows benefit from the section-level construction approach: composers can build a track section by section, matching scene timing, and then export stems for the mix.
Coverage of the August 2025 launch focused primarily on the licensing deals rather than the output quality. TechCrunch reported the music was "cleared for commercial use" and noted the contrast with rivals facing RIAA litigation.[1] Music Business Worldwide called it a significant precedent for the AI music industry, noting that ElevenLabs was the first major AI music entrant to secure deals with both a recordings organization (Merlin) and a publishing organization (Kobalt) simultaneously.[3]
Kobalt CEO Laurent Hubert stated that ElevenLabs "has been a collaborative partner, committed to sourcing data directly from rightsholders," a comment that industry observers noted as unusual given the adversarial tone of most AI-music industry relationships at the time.[4]
The Eleven Album in January 2026 received attention in Variety, Billboard, Adweek, and NBC News, primarily for the participation of Minnelli and Garfunkel.[11][12] Coverage treated the project as a signal that some established artists viewed AI music tools as creatively viable rather than purely threatening.
The ElevenMusic iOS app launch in April 2026 drew comparisons to Spotify in its discovery features alongside Suno and Udio in its creation tools.[6][7] TechCrunch noted that ElevenLabs was positioning the app to compete with both AI music generators and streaming services simultaneously.[6] Music Business Worldwide described it as "taking on Suno and Udio on mobile."[7]
The App Store reception was mixed in early reviews. One experienced producer praised the AI music quality as having "genuinely blown my mind," while other users reported technical issues including difficulty downloading generated songs and app freezing on initial versions.[14]
Independent testing in comparative reviews placed ElevenLabs third in overall output quality behind Suno and Udio for most genres, but first for commercial licensing safety.[15][16] Reviewers consistently recommended Eleven Music for YouTube monetization, advertising, and client work where legal clarity takes priority, and recommended Suno and Udio for higher-volume personal or creative projects where output quality per dollar was more important.[15][16]
The Music v2 launch in late May 2026 drew coverage focused on its editing and genre-transition features.[20][21] TechCrunch led with the model's ability to switch genres mid-track, citing the opera-to-metal example, and noted that the release came nearly ten months after the first version.[20] The Decoder framed the genre transitions as the standout claim and reported the accompanying Believe licensing deal.[21] Several outlets highlighted inpainting as the most practically useful addition, comparing it to non-destructive editing in a digital audio workstation, and noted that the price cuts lowered the cost of API-based music generation.[20][23] Coverage continued to treat licensed training data and commercial clearance as the platform's defining advantage over Suno and Udio.[23][24]
Several limitations of Eleven Music have been noted by independent testers and reviewers.
Vocal quality was the most consistent criticism of the v1 model. In a six-month comparative study published in 2026, only 3 of 20 vocal track requests produced usable results, with the majority described as emotionally flat or robotic.[15] Suno's vocal synthesis was rated more natural by the same testers.[15] ElevenLabs says the Music v2 model improves vocals, instrumentation, and arrangement across genres, though independent testing of v2 against this earlier baseline was still limited at release.[19]
The per-minute pricing model is less intuitive for users accustomed to the per-song billing of Suno and Udio.[15] At the Creator plan level, 62 minutes of music equates to roughly 12 to 20 complete songs depending on length, compared to Suno's 500 songs per month at a lower price.[15] The May 2026 price cuts narrowed but did not eliminate this difference.[20] Users with variable music needs may find the credit system better suited to their workflows than those who need a predictable number of songs each month.
The Eleven Music Pro model, announced at launch as a forthcoming tier trained specifically on Merlin and Kobalt catalogs, had not been released as of May 2026.[2] The original announcement described it as arriving "in the coming weeks or months" from August 2025.[2] The Music v2 release, while a major model update, was presented as a general second-generation model rather than the previously described Pro tier.[19] Custom finetunes for v2 were also described as forthcoming at the v2 launch, with finetune support initially limited to v1.[19]
Commercial use is restricted in certain high-profile contexts without prior written approval from ElevenLabs. Users on standard plans cannot use generated audio in television programs, theatrical films, commercial video games, or radio advertisements without clearing the use separately.[13]
Eligibility for the Merlin and Kobalt opt-in royalty programs is limited. Only artists and songwriters who hold 100% of both master recording and publishing rights qualify.[2] This excludes most major-label artists and many independent artists who have partial rights splits with publishers or distributors.
The platform currently does not offer an Android version of the ElevenMusic app, putting it at a disadvantage relative to Suno and Udio, which support both iOS and Android.[7]