ElevenLabs launches two real-time speech-to-speech models, Eleven Realtime

ElevenLabs released two speech-to-speech models, Eleven Realtime. The larger v1 Max scores 98.8% on the Big Bench Audio benchmark with a 0.57-second response time, and the company says no model on the Artificial Analysis leaderboard is both faster and more accurate. The smaller v1 Mini targets low-cost natural conversation and can tell whether the other party has finished speaking or is just saying backchannel fillers like "uh-huh." v1 Max targets more complex tasks such as customer service. Both models are callable through the ElevenAgents platform and the OpenAI-compatible Realtime API, billed per minute plus token fees, with v1 Max priced higher than v1 Mini. For developers building real-time voice dialogue, the OpenAI-compatible interface means existing Realtime API call code can switch with few changes. For model choice, latency- and accuracy-sensitive customer service use cases can use v1 Max, while cost-sensitive conversational products that need natural turn-taking detection can use v1 Mini.