Model Terminal

PersonaPlex-7B-v1

PersonaPlex-7B-v1 is NVIDIA Research's real-time, full-duplex speech-to-speech conversational model that listens and speaks simultaneously using a unified dual-stream Transformer architecture. It supports controllable persona and voice style with low-latency spoken responses, enabling natural turn-taking and interruption handling without a cascaded ASR→LLM→TTS pipeline. Source: supabase.

Value score
Not scored
Context
tokens
Max output
tokens
Price

Capability radar

Peer value bars

Identity

Developer
NVIDIA
Openness
Modalities
Release
Knowledge cutoff
Deprecation
API docs

Benchmarks

No benchmark scores yet.

Compare nearby

Pricing

Input / 1M
Output / 1M
Speed