ElevenLabs Launches Eleven v4 and v4 Turbo with Natural Voice Direction Tags

ElevenLabs launched Eleven v4 and v4 Turbo on September 28 with 90+ languages and audio tags, offering Creator plan promotions and discounted API rates through

tau · October 3, 2026

#ElevenLabs #Eleven-v4 #TTS #AI-Voice #ElevenCreative

ElevenLabs Launches Eleven v4 and v4 Turbo with Natural Voice Direction Tags

ElevenLabs officially launched its next-generation speech synthesis model, Eleven v4, alongside the real-time low-latency variant Eleven v4 Turbo on September 28, 2026. The company positions both models as its most expressive voice releases yet, claiming a top ranking on Artificial Analysis's voice leaderboard upon release, with immediate availability across ElevenCreative, ElevenAgents, and ElevenAPI.

ElevenLabs Eleven v4 speech model announcement showing expressive voice tags and audio interface

Image source: @ElevenCreative / @dee_naliaks

Succeeding the previous v3 generation, the release adopts a new architecture designed to follow acting and directorial cues more faithfully while preserving voice consistency across long-form productions. Alongside the launch, ElevenLabs introduced limited-time promotional pricing and credit fee waivers running through October 12, 2026.

Flagship Eleven v4: Natural Language Voice Direction and 90+ Languages

Eleven v4 (eleven_v4) serves as the company's flagship high-quality model targeted at audiobooks, character performances, voiceovers, and dubbing pipelines.

  • Expanded language coverage: Supports more than 90 languages, expanding from the 70+ supported in v3.
  • Natural language direction tags: Creators can insert plain-text cues such as [laughs], [whispers], or [door slams] directly into scripts to modulate emotional delivery, pacing, and sound effects. The company states that the model follows multi-tag direction sequences more reliably than v3. SSML break tags are disabled in favor of inline natural language formatting.
  • 10,000 characters per request: Handles large generation blocks in a single request, with narrator and character timbres maintaining consistency across extended batches.
  • Faster voice cloning: Built on the updated architecture, the system clones voices with fewer sample requirements; industry reporting notes usable clones generated from as little as 10 seconds of source audio.

Standard API pricing for Eleven v4 is set at $80 per million characters, positioning it above competitors such as Cartesia Sonic 3.6 ($49) and Google's Gemini 3.8 Flash TTS (~$16.49), according to data compiled by Artificial Analysis.

Eleven v4 Turbo for Real-Time Workloads: 100–150ms Latency

Eleven v4 Turbo (eleven_v4_turbo) is engineered specifically for interactive voice bots, conversational agents, and real-time customer engagement pipelines.

  • The official company announcement cites a median inference latency of approximately 100 ms.
  • Independent reporting across tech outlets (such as the-decoder and Brief AI) notes turn-around response latencies around 150 ms, indicating that real-world performance may vary depending on network and client implementation layers.
  • Standard API pricing is $40 per million characters, half the cost of the flagship v4 model.

ElevenLabs highlights potential use cases across customer support, outbound sales, and automated scheduling, though these marketing performance claims should be evaluated independently within production environments.

Promotional Offer Through October 12: v4 Free Allowance and Discounted API Rates

A two-week launch promotion runs through October 12, 2026, after which standard pricing and credit deductions resume.

  • ElevenCreative credit waiver: Subscribers on the $22/month Creator plan or higher can generate with flagship Eleven v4 in ElevenCreative at no credit cost, capped at twice their monthly credit allotment. This promotion does not apply to free-tier accounts and excludes the real-time Eleven v4 Turbo model.
  • Discounted API pricing: During the promotion window, API generation drops to $22 per million characters for Eleven v4 (down from $80) and $11 per million characters for Eleven v4 Turbo (down from $40).
  • Free-tier evaluation: Industry coverage from MindStudio notes that free-tier users can sample v4's new natural language direction features, though this trial access operates under standard free allowances and remains distinct from the 2x Creator plan waiver.

For content teams and audio production pipelines, the launch window offers a substantial cost reduction for processing high-volume voiceovers or multilingual dubbing jobs. However, teams should plan for the post-promotion baseline ($80/$40 per 1M characters) and note that the credit waiver applies exclusively to the flagship v4 variant.

Sources