Home › What’s New › ElevenCreative
BIG NEWS · SEPTEMBER 2026
ElevenCreative: what it is, what it costs, what’s honest
TL;DR verdict
ElevenLabs isn’t just voice anymore. ElevenCreative is the new creator product line — AI video, images, music, and sound effects alongside the voices — all on one subscription, all on one shared credit pool.
The honest math: video needs a paid plan (Starter, $6/month) and burns fast — the $22/month Creator tier covers roughly 705 seconds of video OR 134 minutes of music, if nothing else is used. The video models are brokered, not built (Veo 3.1, Kling, Seedance, Runway), and Image & Video is still labeled beta.
Our affiliate link covers all of it at the same 22% recurring commission — no separate signup, no new link. And our testing is expanding to cover it: anything below we haven’t run our own tests on yet is labeled as such.
Below: what each piece actually does, the real credit math, ElevenAgents and Reception, and what’s on our test roadmap.
Last checked: September 27, 2026.
What ElevenCreative actually is
First, the naming, because it’s confusing on purpose everywhere else: ElevenCreative is a product line, not a rebrand of elevenlabs.io and not a separate company. The site now presents three lines side by side:
- ElevenCreative — the creator suite: AI video, images, music, SFX, and the Studio editor.
- ElevenAgents — the conversational AI platform for voice/chat agents (plus the Reception product below).
- ElevenAPI — the developer platform: TTS, STT, cloning, dubbing, music, SFX, and now image & video generation through one SDK.
The named structure dates to around February 2026 (announced alongside the $500M Series D); its technical foundation was the November 2025 “Image & Video” beta launch.
The core architecture fact to understand: ElevenLabs trains its own audio models (speech, music, SFX) but brokers third-party video and image models inside one credit pool. The pitch is “one subscription, no separate tool subscriptions.” That’s genuinely convenient — and it means video quality comparisons are really “same model, different wrapper plus the audio stack,” not a proprietary ElevenLabs video model beating anyone.
Video generation
Text-to-video and image-to-video, running on brokered third-party models. Confirmed in the product in September 2026: Google Veo 3.1 (Standard/Fast/Lite), Kling 3.0 / 3.3 Pro, Seedance 2.5, Wan 2.5, MiniMax H3 Max, Runway Gen-4.5 / Gen-4 Turbo, LTX 2 Pro, and OmniHuman (avatar/lip-sync engine).
Sora is dead — don’t expect it. OpenAI shut down the Sora Videos API and all model aliases on September 24, 2026. Any source still listing Sora inside ElevenLabs is stale.
Per-plan video allowances (cheapest settings, per third-party pricing trackers — directional until we verify in-app):
- Free: no video at all
- Starter ($6/mo): ~211 seconds/month
- Creator ($22/mo): ~705 seconds/month
- Pro ($99/mo): ~3,540 seconds/month
- Scale ($299/mo): ~11,752 seconds/month
- Business ($990/mo): ~47,008 seconds/month
Honest caveats: Image & Video was still labeled beta as of mid-September 2026. “Upscale to 4K” is a gated feature, reported at Pro ($99+) and above — we could not verify a full per-plan resolution matrix, so treat 4K claims as gated-and-untested. Per-generation credit costs are shown at checkout, not published. Max clip length is model-dependent (typically 4–10 seconds).
Images
AI image generation is included on every plan, including Free (40/month on Free, watermarked, personal use with attribution). The confirmed models are premium third-party ones you’d otherwise buy separately: Nano Banana Pro (Google’s image model) and GPT Image 2.5 (OpenAI).
The catch is credit burn: Nano Banana Pro at 4K costs about 1,818 credits per image — roughly 66 such images a month on the Creator plan. Nano Banana Pro outputs also carry SynthID watermarking. Quotas scale with plan (Starter ~198, Creator ~660, Pro ~3,300, Scale ~11,000, Business ~44,000 per month, per third-party trackers).
Music — ElevenMusic
The current model is Music v2.5 (September 2026), with the biggest gains in vocal-led and acoustic genres — R&B, soul, hip hop, rock, metal, orchestral, cinematic. Specs: text-to-music with vocals, 3 seconds to 5 minutes per generation, 44.1 kHz, MP3/WAV, timed lyrics, section inpainting, up to 6 stems separation, and a force-instrumental option.
Cost: about 900 credits per minute in-app — roughly 134 minutes a month on Creator if nothing else is used. API pricing is $0.15/min.
The licensing story is the strongest in AI music, and it’s real: ElevenMusic is trained on licensed data (Merlin, Kobalt — not major-label catalogs), and commercial use starts at the $6/month Starter tier. Two honest limits: self-serve plans exclude film, TV, and studio games (that clearance is Enterprise-only), and third-party editorial rates its vocal quality below Suno’s — we haven’t run our own music A/B yet, so we’ll publish when we have.
Music Marketplace (launched March 2026): creators publish AI tracks; others download, remix, or license them across three tiers (Social Media / Paid Marketing / Offline). ElevenLabs reports $11M distributed to creators. The honest caveat: no exclusivity — a similar prompt can yield a near-identical track, and the legal risk sits with the user.
The UMG deal (announced September 10, 2026) is a multi-year licensing and product-development agreement — ElevenLabs’ first with a major label. They’re building a separate fan platform for licensed remixes and mashups from opt-in UMG artists, distinct from ElevenMusic. As of late September 2026 there is no launch date, no pricing, and no artist lineup — treat it as a watch item, not a feature.
Sound effects
Sound Effects v2 does text-to-sound at 44.1/48 kHz, up to 30 seconds per generation, with seamless loop support, 38 built-in categories plus open prompting. It’s rated close to studio quality in 2026 roundups, and it’s included on every plan including Free.
The honest limit: the 30-second cap. Long ambient beds need stitching — plan for it.
Lip-sync & Studio
Lip-sync is offered as an in-flow step on all paid plans — sync a talking head to any ElevenLabs voice, including clones, using Veed or OmniHuman engines. It’s credit-hungry (reviewers note it’s where most of the credits go), and sync quality follows whichever underlying video model you pick. We haven’t benchmarked it against dedicated lip-sync tools yet.
Studio 4.0 (September 2026) puts video and image generation directly on a multi-track timeline — frame-level zoom, clip snapping, captions as editable timeline clips, and team comments/review. Studio Agent (May 2026) is the conversational co-editor: it drafts scripts, places clips, and generates voiceovers — web-only, no API, billed on token use. Project limits: Free gets 3 projects, Starter gets 20. Audio export is 128 kbps MP3 on lower tiers, 16-bit 44.1 kHz WAV or 192 kbps MP3 on Pro and above.
Flows is the node-based automation canvas for repeatable generation chains, with real-time collaboration added in May 2026. Developers get the Image & Video API (18 models, beta, gated at Pro and above) and a Creative MCP that puts 50+ models inside Claude, ChatGPT, and Cursor.
The honest credit math
Everything — voice, video, image, music, SFX, dubbing, Scribe, API use — draws from one shared credit pool. No separate creative add-on fees, which is the convenience. The trap: one heavy video or dubbing project can eat the month’s voice budget. And failed or glitched generations still consume credits — documented in user complaints — so budget a retry margin.
- Free ($0): 10,000 credits — non-commercial, watermarked, attribution; image/music/SFX but no video; 3 Studio projects.
- Starter ($6/mo): 30,000 credits — commercial license; Instant Voice Cloning; video (~211 s); music downloads; Dubbing Studio; 20 Studio projects.
- Creator ($22/mo, $11 first month): 121,000 credits — Professional Voice Cloning; video (~705 s) OR ~134 min of music OR ~121 min of narration — pick one, they share the pool.
- Pro ($99/mo): 600,000 credits — 4K upscale; Image & Video API access; video (~3,540 s).
- Scale ($299/mo): 1,800,000 credits — team seats and collaboration.
- Business ($990/mo): 6,000,000 credits — video (~47,008 s).
- Enterprise: custom — SSO, HIPAA, SLAs, film/TV rights.
Annual billing runs about 17% off, and paid plans roll over unused credits while subscribed. When the free tier is enough: non-commercial experiments — learning the tools, testing prompts, hobby projects. The week anything earns money, Starter’s $6 commercial license is the honest entry point.
Watch out for stale third-party pages showing old figures ($5 Starter, $330 Scale, $1,320 Business) — the table above reflects the current lineup as of September 2026. Always verify at elevenlabs.io/pricing before buying.
ElevenAgents & Reception
Beyond the creator suite, ElevenAgents is ElevenLabs’ no/low-code platform for building conversational voice, chat, SMS, and email agents — customer support, lead qualification, appointment booking. It’s billed separately from the shared credit pool, per minute: about $0.08/min standard ($0.16/min burst to exceed concurrency caps), with LLM and telephony billed on top at cost. Included minutes run from 15 (Free) to 12,375 (Business); unlimited agent count, but concurrency caps are the binding constraint. Enterprise adopters cited include NVIDIA, Deutsche Telekom, Salesforce, and Santander.
Reception (launched September 16, 2026) is the packaged small-business version — a 24/7 AI phone receptionist with appointment booking, call transfer, and message-taking. Basic is $29/month (75 credits, 1 receptionist, 1 number) with a 14-day free trial; Plus $79 and Premium $199 tiers above it. Honest limits: no HIPAA compliance (healthcare is directed to enterprise), and AI-call disclosure is legally required in several US states and under the EU AI Act.
What we’ve tested vs. what’s next
Voice is where our testing lives today: Test 01 (Brian vs. Sarah, same script, identical settings) is published with playable samples in our test library, and more voice protocols are written and queued.
Video, music, SFX, and images are not yet tested by us. This page is the product map — the honest lay of the land from published pricing and documentation. Hands-on tests publish here with samples, settings, and dates, like everything else on this site. Planned tests stay labeled “planned” until the media exists.
On our roadmap, in the order we think they matter:
- The “photorealistic 4K” blind test — ElevenCreative outputs vs. real footage, judged blind.
- Wrapper vs. direct — the same model (Veo, Kling) through ElevenLabs vs. direct: is the convenience worth the credit premium?
- The credit-burn receipt — one 8-second clip plus voiceover, music, and SFX, with every credit accounted for.
- Lip-sync quality — against dedicated lip-sync tools.
- Music Content ID test — ElevenMusic vs. competitors on claim rates.
FAQ
Is ElevenCreative a separate product from ElevenLabs?
No. It’s a product line inside ElevenLabs — not a rebrand, not a separate company. elevenlabs.io now presents three lines: ElevenCreative (creators), ElevenAgents (conversational AI), and ElevenAPI (developers), all on one subscription.
Does the ElevenLabs affiliate program cover ElevenCreative?
Yes. The Creator Affiliate Program pays 22% recurring commission for the first 12 months on the subscription plan — and since all three product lines run on that plan, one affiliate link covers everything. Ours is disclosed next to every link on this site.
Can I generate AI video on ElevenLabs’ free plan?
No. Video needs a paid plan starting at Starter ($6/month, ~211 seconds at the cheapest settings). The free tier includes image, music, and SFX generation but no video and no commercial rights.
Is ElevenLabs’ AI video actually 4K?
“Upscale to 4K” is gated at the Pro tier ($99/month) and above, and Image & Video was still labeled beta as of mid-September 2026. We haven’t run our own resolution tests yet — treat photorealism claims as untested until independent tests publish.
Is ElevenLabs’ AI music safe for monetized YouTube videos?
ElevenMusic’s licensed training data and commercial use from the $6 Starter tier make it the strongest licensing story in AI music. But self-serve plans exclude film, TV, and studio games — that’s Enterprise-only. We haven’t run our own Content ID claim-rate test yet.
What’s the catch with one shared credit pool?
One heavy video or dubbing project can eat the month’s voice budget, since everything draws from the same credits. Failed generations still burn credits, so budget a retry margin. Creator ($22/month) covers roughly 705 seconds of video or 134 minutes of music — if nothing else is used.