Play 3.0-mini
PlayAI's conversational voice model: sub-second streaming TTS in 30+ languages with instant cloning — the low-latency pick for real-time voice agents.
About
Play 3.0-mini is PlayAI's (formerly PlayHT) flagship conversational voice model, positioned squarely at real-time voice agents rather than studio narration. The pitch is latency: the company claims under 130ms time-to-first-byte, and third-party voice-stack guides consistently put PlayAI's streaming output in the 200–600ms practical range, noticeably under ElevenLabs' non-flash models. It covers 30+ languages, offers 200+ prebuilt voices, and supports instant voice cloning from short samples, with the voiceEngine parameter (Play3.0-mini) selecting it in the streaming API. PlayAI exposes REST and streaming endpoints with Python and Node SDKs, and has expanded the platform into AI voice agents, a multi-turn PlayDialog model, and the PlayNote podcast tool. Pricing is usage-based with a reported 5,000 characters/month free tier — the pricing page itself leads with agent plans, so confirm current rates before committing. Quality has closed the gap with ElevenLabs in early-2026 comparisons, though English expressiveness still trails the best studio models.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
- Provider-claimed <130 ms time-to-first-byte; third-party guides report 200–600 ms practical streaming latency
- G2 rating 4.3/5 (third-party); no independent latency benchmarks found
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 6/10
Scored 6/10 (TIER metrics, metrics-v1). Evidence: widely integrated platform (200+ prebuilt voices, major voice-agent stacks); rebrand from PlayHT to PlayAI with agent products (PlayDialog, PlayNote). Strengths: genuinely low-latency streaming, 30+ languages, instant cloning, developer-friendly API surface. Docked for: (1) closed weights, all latency and quality claims are provider-claimed with third-party reports showing practical latency above the <130ms headline; (2) free-tier and plan details are murky on the public pricing page; (3) launch date never published precisely — 3.0 series current through 2026. License: proprietary.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 8.0/10
- Openness · license + docs
- 2.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · PlayAI (rebranded from PlayHT); Play 3.0-mini is the conversational flagship — voiceEngine=Play3.0-mini
- · 30+ languages, 200+ prebuilt voices, instant cloning; provider-claimed <130ms TTFB
- · Third-party guides report 200-600ms practical streaming latency — above the headline
- · Pricing page leads with agent plans; 5,000 chars/month free tier is third-party-reported; launch date approximate