Legacy model notice — verified August 20, 2026: MiniMax’s current pay-as-you-go table explicitly labels Music 2.0 discontinued, and the current Music Generation request enum does not include
music-2.0. The separate rate-limit table still names Music 2.0, but that capacity row is not evidence of account entitlement or request acceptance. Do not use Music 2.0 for a new integration. New users are directed to MiniMax Audio or the downloadable MiniMax Music 3 checkpoint; eligible existing paying API users should verify their console before relying on any hosted music model.
MiniMax Music 2.0 was an important step in MiniMax’s music-generation line. It focused on expressive vocals, memorable melodies, controlled instrumentation, and complete songs lasting up to five minutes. This page preserves the model’s documented history without presenting it as a safe choice for a new integration.
The distinction matters because MiniMax’s official pages are not perfectly synchronized. The model catalog and pricing page still preserve Music 2.0 as a legacy model, and the rate-limit table still names it. However, the live Music Generation reference defines the request field model as an enum, and Music 2.0 is absent from that operational list. When documentation conflicts, the endpoint reference and accepted model list are the stronger sources for implementation decisions.
Independent reference: MiniMax-AI.chat is not affiliated with MiniMax and cannot confirm private or grandfathered model access. Only MiniMax can confirm whether a specific account still has a supported Music 2.0 route.
Contents
- Music 2.0 status at a glance
- The October 29 and October 31 date conflict
- What Music 2.0 was designed to do
- What the current API documentation means
- Music 2.0 vs Music 2.6 vs Music 3.0
- Migration guidance
- Legacy pricing and rate limits
- Rights and publishing cautions
- Frequently asked questions
- Official sources
Music 2.0 status at a glance
| Question | Verified answer |
|---|---|
| Is Music 2.0 the recommended MiniMax music model? | No. MiniMax lists music-3.0 as recommended. |
| How does MiniMax classify Music 2.0? | Legacy text-to-music model. |
Is music-2.0 in the current Music Generation request enum? | No. |
| Does MiniMax still publish a price for it? | Yes: $0.03 per up-to-5 minutes of music, in the legacy pricing table. |
| Does the rate-limit page still name it? | Yes: the shared music table shows 120 RPM and 20 concurrent connections. |
Should a new integration send music-2.0? | No, not without direct confirmation from MiniMax for that account. |
| Best default for a new build | MiniMax Audio or the downloadable MiniMaxAI/MiniMax-Music3 checkpoint; do not plan a new hosted music API integration. |
This page is therefore a historical product guide, not a copy-and-paste API tutorial. It intentionally contains no Music 2.0 request example. A code block using an identifier omitted from the endpoint’s accepted values would create false confidence and could fail even though older pricing and quota tables remain online.
The October 29 and October 31 date conflict
Two first-party MiniMax pages give different dates for Music 2.0:
- The official model release notes record October 29, 2025.
- The official launch story, “Making Music Creation Accessible to Everyone,” is dated October 31, 2025.
The safest editorial treatment is to preserve both dates rather than silently choosing one. October 29 is MiniMax’s catalog release record; October 31 is the date displayed on the detailed product announcement. The two-day difference may reflect a staged rollout, documentation timing, or later publication of the launch article, but MiniMax does not explain the discrepancy. This guide does not infer a reason.
For timelines and structured data, use wording such as “recorded in MiniMax release notes on October 29, 2025; detailed launch article published October 31, 2025.” Avoid presenting either date as independently verified real-world availability in every region or interface.
What MiniMax Music 2.0 was designed to do
MiniMax’s current model directory summarizes Music 2.0 as text-to-music generation with enhanced musicality, natural vocals, smooth melodies, human-like performance, richer emotional expression, and stronger tone control. Its launch article gives more detail. The following capabilities are historical first-party product claims, not results from an independent benchmark performed by this site.
Expressive vocals and varied singing styles
Music 2.0 was positioned around more human-like vocal texture and closer control of phrasing, breath, rhythm, and emotion. MiniMax demonstrated prompts that retained a core vocal identity while changing style, and it named pop, jazz, blues, rock, and folk among the supported directions. The announcement also showcased male-female duets and a cappella arrangements.
These examples describe the intended range of the model; they are not guarantees that every prompt will preserve identity, pronounce every lyric correctly, or reproduce a requested performance consistently. Generated vocals should be reviewed for intelligibility, unintended resemblance, artifacts, and suitability for the intended audience.
Complete song structure and memorable melodies
The launch material said Music 2.0 could generate structurally complete songs with verses, choruses, and bridges, with a possible duration of up to five minutes. MiniMax also emphasized hooks that were easier to remember and melodies intended to follow familiar compositional patterns.
“Up to five minutes” is a maximum product description, not a promise that every output will use the full duration. Structure, repetition, pacing, and ending quality can vary, so production workflows should evaluate the entire audio file rather than judging only an opening excerpt.
Instrument direction and fuller audio
MiniMax said users could direct individual instruments and build layered arrangements with a natural groove. A jazz example described saxophone, trombone, trumpet, drums, and piano entering in sequence. The company also presented Music 2.0 as an improvement in vocal texture, instrument placement, and overall spatial presence.
The announcement extended that expressive positioning to film-style monologues and poetry recitation with music. Prompts could describe vocal emotion and soundscape to shape a staged emotional progression. This made the model relevant not only to conventional songs but also to narrative audio experiments, trailers, spoken-word pieces, and concept demos.
What the current API documentation means
MiniMax’s documentation currently sends three different signals:
| Official page | What it says about Music 2.0 | How to interpret it |
|---|---|---|
| Model directory | Lists Music 2.0 under Legacy Models. | Confirms historical product status and feature summary. |
| Pay-as-you-go pricing | Lists Music 2.0 as discontinued and preserves the $0.03 historical unit price. | Preserves a legacy commercial entry, but does not prove request acceptance. |
| Rate limits | Includes Music 2.0 in a music table with 120 RPM and 20 CONN. | Preserves a capacity entry, but does not establish that new Music 2.0 requests are accepted. |
| Music Generation API reference | Omits music-2.0 from the required model enum. | Strongest evidence for what developers should send in a new request. |
MiniMax does not explain why the legacy pricing and rate-limit entries remain while the operational model enum omits Music 2.0. Do not infer availability from those entries.
If an existing application still uses Music 2.0 successfully, do not assume that success guarantees indefinite support. Record the account, endpoint, model identifier, response metadata, and date of the last successful call. Ask MiniMax support for an explicit deprecation or migration position before making business commitments around it. New applications should follow the live enum and the current API guide.
MiniMax Music 2.0 vs Music 2.6 vs Music 3.0
| Area | Music 2.0 | Music 2.6 | Music 3.0 |
|---|---|---|---|
| Documentation status | Legacy | Previous-generation model; hosted access only for eligible existing paying accounts | Recommended model; hosted access closed to new users |
| Current paid model ID | Not in live enum | music-2.6 | music-3.0 |
| Free API variant | Not listed | Discontinued | Discontinued |
| Official positioning | Enhanced musicality, natural vocals, smooth melodies, emotional expression, tone control | “Cover Reborn. Bass Redefined.” | Intent understood, sound elevated, vocals humanized |
| Published paid price | $0.03 per up-to-5 minutes of music | $0.15 per up-to-5 minutes of music | $0.15 per up-to-5 minutes of music |
| Published paid RPM | 120 in rate-limit table, despite enum omission | 120 | 120 |
| Best use now | Historical research or documented legacy maintenance | Maintaining an entitled existing 2.6 workflow | Eligible existing-account migration, MiniMax Audio, or open-weight Music 3 evaluation |
The comparison is intentionally based on MiniMax’s own documentation rather than an invented quality score. MiniMax does not publish a controlled 2.0-versus-2.6-versus-3.0 listening benchmark on the referenced API pages. A fair independent comparison would need the same lyrics, prompt, audio settings, multiple generations, blind evaluation, and a disclosed rubric for vocals, structure, prompt adherence, production quality, and artifacts.
For broader market context, see our MiniMax Music vs Suno vs Udio comparison. That page should be treated separately from this version-specific historical record.
Migration guidance for Music 2.0 projects
Migration is not just a model-name substitution. Music models can interpret the same direction differently, so changing from 2.0 to 2.6 or 3.0 can affect vocal delivery, section timing, instrumentation, loudness, and the balance between prompt and lyrics.
- Inventory the dependency. Find every hard-coded model ID, configuration file, job template, test fixture, dashboard, and billing assumption related to Music 2.0.
- Preserve representative inputs. Save a small, rights-cleared test set covering short and long lyrics, instrumental directions, multiple genres, duets, difficult pronunciation, and edge cases.
- Choose the target deliberately. Use Music 3.0 for new work. Keep Music 2.6 only when continuity with a validated 2.6 pipeline matters.
- Revalidate request fields. Follow the current Music Generation reference for prompt and lyrics requirements, supported structure tags, instrumental mode, lyric optimization, output format, and audio settings.
- Compare outputs, not marketing labels. Review full tracks for pronunciation, unwanted resemblance, section transitions, clipping, silence, endings, and consistency across repeated generations.
- Update cost and capacity controls. Moving from the published Music 2.0 legacy price of $0.03 to a $0.15 current model changes the listed unit cost by five times.
- Roll out gradually. Keep logs and trace IDs, monitor failures, and maintain a reversible application configuration while the new model is evaluated.
Do not build a fallback that silently sends Music 2.0 after a current model fails. A fallback should use another identifier explicitly allowed by the live reference and should preserve user expectations about price, quality, and output behavior.
Legacy pricing and rate limits
MiniMax’s pay-as-you-go page labels Music 2.0 discontinued while preserving a historical price of $0.03 per up-to-5 minutes of music. The separate rate-limit guide still shows 120 RPM and 20 concurrent connections, but that row is not proof of availability.
These are documented figures, not an offer from this independent site and not an availability guarantee. Taxes, account entitlements, contracts, regional access, negotiated limits, and documentation changes may alter what a user can actually purchase or call. Confirm the live console and written MiniMax terms before estimating production costs. Our MiniMax pricing guide summarizes the broader API and Token Plan structure.
A simple historical estimate illustrates the published pricing difference: 1,000 legacy Music 2.0 billing units at $0.03 would equal $30, while 1,000 paid Music 2.6 or Music 3.0 billing units at $0.15 would equal $150. This arithmetic excludes retries, lyric-generation charges, taxes, subscriptions, negotiated terms, and any credits. It does not establish that new Music 2.0 requests or billing units can still be purchased.
Rights, voice, and publishing cautions
AI music generation creates legal and editorial questions that model documentation alone does not answer. Before publishing or monetizing a track, confirm that you have permission to use every protected input: lyrics, melodies, reference recordings, samples, names, trademarks, and any identifiable voice or performance characteristics. Do not ask for impersonation or use a person’s voice, identity, or work in a way that violates consent, publicity, copyright, contract, or platform rules.
A generated file is not a warranty that the result is original, non-infringing, commercially usable, or eligible for copyright protection. Listen for accidental similarity, document the prompts and source materials, keep generation records, and obtain qualified legal review for high-value releases, advertising, client work, or distribution across multiple territories.
Review the MiniMax Audio Music Creation Terms of Use and the API Platform Terms for the exact account and product used. The music-specific terms require users to hold the necessary rights for inputs, obtain authorization for submitted voice data, use the feature lawfully, and clearly label AI-generated output to avoid misleading others. They do not provide a blanket commercial-rights guarantee. This section is general risk guidance, not legal advice.
Frequently asked questions
Is MiniMax Music 2.0 still available through the API?
No. The current pay-as-you-go table explicitly labels Music 2.0 discontinued, and the Music Generation model enum does not include music-2.0. The separate rate-limit row is not proof of availability. Ask MiniMax directly only when maintaining a pre-existing dependency.
Why is there no Music 2.0 API example on this page?
A publish-ready example should use a model identifier accepted by the live endpoint reference. Providing a copyable Music 2.0 request while the enum omits that identifier would be misleading. Use the official Music Generation reference and the Music 3.0 guide instead.
When was Music 2.0 released?
MiniMax’s release notes record October 29, 2025, while the detailed launch article is dated October 31, 2025. Both dates are first-party records. This page reports the discrepancy rather than inventing an explanation.
How long could a Music 2.0 song be?
MiniMax’s launch article and legacy pricing entry describe songs of up to five minutes. “Up to” is a maximum, not a guaranteed output duration.
Was Music 2.0 limited to vocals?
No. Its launch material emphasized vocals, but it also described controllable instrumentation, layered arrangements, complete song structures, monologue soundtracks, and poetry recitation with music.
Should I use Music 2.6 or Music 3.0?
Do not start a new hosted integration with Music 2.6 or Music 3.0: MiniMax closed paid Music API access to new users on August 20, 2026. New users should evaluate MiniMax Audio or the separate open-weight Music 3 checkpoint. An eligible existing paying account may still compare Music 2.6 and Music 3.0 for a maintained workflow.
Is Music 2.0 cheaper?
Its discontinued legacy listing preserves a historical price of $0.03 per up-to-5 minutes of music, compared with $0.15 per up-to-5 minutes of music for paid Music 2.6 and Music 3.0. The lower historical listing is not useful for a new cost plan unless MiniMax confirms that the model remains available to the relevant account.
Can I commercially release music generated with MiniMax?
Do not answer that question from a model guide alone. Review the current terms for the exact product and account, clear all input rights, assess the output, and obtain legal advice when the release carries meaningful commercial or reputational risk.
Official sources and verification notes
- MiniMax Music 2.0 launch article — dated October 31, 2025; source for the historical vocal, melody, instrument, song-length, and narrative-audio claims.
- MiniMax model release notes — records Music 2.0 on October 29, 2025, Music 2.6 in April 2026, and Music 3.0 on July 16, 2026.
- MiniMax model directory — classifies Music 2.0 as legacy and summarizes Music 2.6 and Music 3.0.
- Music Generation API reference — source for the current accepted model enum and Music 3.0 recommendation.
- MiniMax pay-as-you-go pricing — preserves the Music 2.0 legacy price and publishes current Music 2.6 and Music 3.0 prices.
- MiniMax rate limits — still includes Music 2.0 in the shared music-capacity table.
- MiniMax Audio Music Creation Terms of Use — covers input rights, submitted voice authorization, lawful use, and AI-output labeling.
Verification method: claims were checked against MiniMax’s live official documentation on July 17, 2026. Where official pages conflict, this guide states the conflict and avoids inferring availability. MiniMax-AI.chat is an independent publisher and is not affiliated with or endorsed by MiniMax.