Eleven v4 and 16 more new image, video, and audio models
Eleven v4 voices, Luma Ray 3.2, LTX-2.5, Lyria 3.5, and newer versions of Grok, Ideogram, Qwen, and PixVerse.
The generators now carry the latest model from most studios, and ElevenLabs speech costs less.
Voice and audio
- Eleven v4 is ElevenLabs' most expressive voice model, with 85 languages and up to 10,000 characters per clip. Eleven v4 Turbo is the faster, cheaper version.
- Gemini 3.8 Flash TTS and Gemini 3.8 Flash Lite TTS replace Gemini 3.1 Flash TTS.
- Eleven Music v2.5 is now the default music model. Lyria 3.5 writes full songs with vocals.
- Mirelo SFX 1.6 generates sound effects from a description.
- Eleven v3, Multilingual v2, and Flash v2.5 now cost fewer credits per clip.
Video
- Luma Ray 3.2 and LTX-2.5 create videos from text or a first frame. LTX-2.5 can add sound.
- PixVerse V6, Kling 3.0 Turbo, and Wan 3.0 Prime join their families.
- Mirage Avatar X turns a photo and a voice track into a talking video.
Images
- Grok Imagine 2.0 and Qwen-Image 3.0 generate and edit images.
- Ideogram 4.0 renders text in images more accurately than 3.0.
- Seedream 5.0 Flash is the cheapest Seedream.
Imagen 4 is no longer available from our provider, so it has been removed. Nano Banana 2 is the closest replacement.
Gen-4 Aleph has also been removed from the video models.