ElevenCreativeElevenAgentsElevenAPIResourcesEnterprisePricing
ElevenCreative
ElevenAgents
ElevenAPI
Resources
Enterprise
Pricing
Sign upSign up
ElevenCreativeElevenAgentsElevenAPIResourcesEnterprisePricing
ElevenCreative
Introduction
Accuracy improvements
Availability
ElevenAgents
ElevenAPI
Resources
Enterprise
Pricing
Sign upSign up
Eleven v3, our most advanced Text to Speech model, is now out of Alpha and generally available.
On this page
Introduction
Accuracy improvements
Availability
Since the Alpha release, we've continued refining the model. Two key improvements:
More stable. In testing, users preferred the new version 72% of the time over the previous Alpha release.
More accurate. We significantly improved how the model handles numbers, symbols, and specialized notation across languages.
Accuracy improvements
Text to Speech models need to interpret what you write and decide how to say it. The same symbols can mean different things in different contexts.
Consider a phone number: "+49 170 9876543"
In some cases, our models would read this as "plus forty-nine, one hundred seventy, nine million eight hundred seventy-six thousand five hundred forty-three" - interpreting the digits as large numbers rather than a digit sequence. The correct reading is "plus four nine, one seven zero, nine eight seven six five four three."
These kinds of errors showed up across categories - sports scores, chemical formulas, currencies, coordinates - anywhere the models had to interpret symbols and decide how to vocalize them.
We tested against an internal benchmark covering 27 categories across 8 languages.
Overall: 68% reduction in errors. Error rate dropped from 15.3% to 4.9%.
Error rate by category:
The improvements are most significant in categories where context determines interpretation - where a colon might indicate a sports score, a time, or an aspect ratio depending on surrounding text.
Examples
Currencies — correct magnitude:
Input: ¥250,000
Before: 25,000 yen
After: 250,000 yen
Chemical formulas — symbols preserved correctly:
Input: SO₂
Before: "sulfur double" (garbled)
After: "S O two"
Sports scores — context-aware interpretation:
ElevenCreativeText to SpeechSpeech to TextVoice ChangerText to Sound EffectsVoice CloningVoice IsolatorAI Music GeneratorStudioVoice DesignAI Voice GeneratorAI Image GeneratorAI Video Generator
Input: Final score: 102-98
Text to Speech
Speech to Text
Voice Changer
Text to Sound Effects
Voice Cloning
Voice Isolator
AI Music Generator
Studio
Voice Design
AI Voice Generator
AI Image Generator
AI Video Generator
ElevenAgentsVoice AgentsConversational AIIntegrationsTelecommunicationsFinancial ServicesHealthcareTechnologyRetail & E-commerceCustomer SupportChatbots
Before: "one hundred two minus ninety-eight"
Voice Agents
Conversational AI
Integrations
Telecommunications
Financial Services
Healthcare
Technology
Retail & E-commerce
Customer Support
Chatbots
ElevenAPIAPI ReferenceAgents APIDubbing APIText to Speech APISpeech to Text APISound Effects APIMusic APIAPI Key
After: "one hundred two to ninety-eight"
API Reference
Agents API
Dubbing API
Text to Speech API
Speech to Text API
Sound Effects API
Music API
API Key
ResourcesBlogIconic MarketplaceImpact ProgramStartup GrantsHelp CenterWebinarsDocsEnterpriseTrust CenterIndia
Availability
Blog
Iconic Marketplace
Impact Program
Startup Grants
Help Center
Webinars
Docs
Trust Center
India
SocialsXX - DevelopersLinkedInGitHubYouTubeYouTube - DevelopersDiscordTikTokInstagramFacebookReddit
Eleven v3 is now generally available across all platforms.
X
X - Developers
GitHub
YouTube
YouTube - Developers
Discord
TikTok
CompanyAboutCareersSafetyBrand & Press KitEU Digital Services Act (DSA)ElevenLabs SummitTermsPrivacyModern Slavery PolicyCCPA NoticeEU-US DPF PolicyAI TransparencyCookie Settings
Explore articles by the ElevenLabs team
About
Careers
Safety
Brand & Press Kit
EU Digital Services Act (DSA)
ElevenLabs Summit
Terms
Privacy
Modern Slavery Policy
CCPA Notice
EU-US DPF Policy
AI Transparency
Cookie Settings
.webp)
