Plain-English Summary
A very fast, lower-cost speech model designed for voice agents and interactive applications.
Eleven Flash v2.5
A very fast, lower-cost speech model designed for voice agents and interactive applications.
Eleven
Audio / Speech
GA
Open weights: No
API: Yes
Technical Notes
Low-latency text-to-speech model with roughly 75ms model latency, 32 languages, and a 40,000-character request limit.
Best For
Voice agents, chatbots, games, real-time applications, bulk speech
Watch Out For
Less expressive than Eleven v3 or Multilingual v2.
Model Specs
- Maker
- ElevenLabs
- Country
- USA
- Family
- Eleven
- Type
- Audio / Speech
- Release date
- Not listed
- Date confidence
- Not publicly stated
- Status
- GA
- Context window
- Varies
- Max output
- Not listed
- Parameters
- Not disclosed
- Active parameters
- Not disclosed
- Architecture
- Proprietary low-latency speech model
- Reasoning
- Not applicable
- Tool calling
- No
- Structured output
- No
- API available
- Yes
- Open weights
- No
- Self-hostable
- No
- License
- Proprietary
- Inputs
- Text
- Outputs
- Audio
- Approx. price
- $0.05 per 1,000 characters
- How to access
- Provider website or API
- Model / API ID
- eleven_flash_v2_5
- Knowledge cutoff
- Not publicly disclosed
- Fine-tuning
- Voice selection and cloning
- Completeness
- 76 %
- Verified
- Jul 18, 2026
- Snapshot date
- Jul 18, 2026
- Source
- Official source