ElevenLabs Launches Eleven v4, Its Fastest and Most Emotive Voice Model Yet

ElevenLabs launched Eleven v4 on September 28, 2026, alongside a low-latency Eleven v4 Turbo variant built for conversational agents. The company calls it its “fastest and most emotive voice model” yet, rolling out immediately across ElevenAgents, ElevenCreative, and the ElevenLabs API.

Details

• Emotional expressiveness: the models read “tone, pacing, emotion, character, and context” to produce speech that sounds dramatic, tender, urgent, comedic, or conversational, rather than flatly narrated
• Inline delivery tags: creators can drop in tags like “[laughs]” or “[said angrily in French accent]” to control specific line readings directly in the script
• More natural multi-speaker dialogue: in multi-voice scenes, speakers now respond to conversational context so exchanges sound like real dialogue instead of separately generated lines stitched together
• Better voice consistency: an improved speaker-identity capture method preserves each voice’s distinct qualities across long-form content
• Faster instant voice cloning: Instant Voice Clones now need only 10 seconds of reference audio, down from ElevenLabs’ prior requirement
• Speed for real-time use: Eleven v4 Turbo posts a median time-to-first-speech of about 150 milliseconds, aimed at low-latency conversational agents
• Broad language coverage: both models support more than 90 languages with improved accent adherence and more native-sounding pronunciation when a voice is used across languages
• Benchmark claims: ElevenLabs says blind listener tests preferred Eleven v4 over competing models roughly 75% of the time, and cites a #1 ranking from Artificial Analysis
• Target use cases: audiobooks, character performances, voiceovers, dubbing/localization, and customer support agents

What happened next

Eleven v4 is available immediately to users of ElevenAgents, ElevenCreative, and the API, with no separate pricing changes announced alongside the launch. The release follows ElevenLabs’ Music v2.5 rollout in September and its broader push to compete on raw voice-model quality against rivals, positioning emotional range and multi-speaker realism as its main differentiators heading into Q4.