Context
—
Max output
—
Input / 1M
—
Output / 1M
—
Modality
Chat
Cutoff
unknown
About this model
Flux is Deepgram's conversational speech recognition model built specifically for voice agents, with model-integrated end-of-turn detection and turn-taking, at Nova-3 level accuracy.
Best suited for
- Voice agent turn-taking
- Real-time conversational transcription
Supported parameters
eager_eot_thresholdEager End-of-Turn ThresholdConfigures how eagerly Flux predicts a speaker has finished their turn, for low-latency voice agent responses.
languageSets the language to use for processing, transcription, or generation.
smart_formatApplies automatic formatting improvements for cleaner final output.
typeIndicates the configuration category or payload type.
Use this model
OpenAI-SDK compatible, with gateway fallback and routing across providers.
Other Deepgram models
| Model | Context | Max output | Input | Output | Capabilities |
|---|---|---|---|---|---|
| nova-3-medical Chat | — | — | — | — | — |
| flux-general-multi Chat | — | — | — | — | — |
| nova-2 Chat | — | — | — | — | — |
| nova-3 Chat | — | — | — | — | — |
Pricing & provider details
Deepgram
deepgram/flux-general-en
- Input · per 1M
- —
- Output · per 1M
- —
- Cached input · per 1M
- —
- Context window
- —
- Max output
- —
- Knowledge cutoff
- unknown
- Auto-router
- Not supported
Start building today
Route flux-general-en — and every other model in the catalogue — through one endpoint, with failover built in.