Phonic-1 is a serious option for teams building phone agents that need to act fast, and the flat per-minute price is refreshingly simple.
AI Agents · Tool review
Phonic AI: Phonic-1 Voice Agents & Pricing
The AI news that actually mattered this week.
One short email every Friday: the biggest AI stories, what they mean for you, and the tools worth trying. A 3-minute read, no hype.
Free. Unsubscribe anytime.
You're in. Watch for the brief on Friday. Free. Unsubscribe anytime.
Summary
Pros and cons
What we like
- One speech-to-speech model, no transcript chain.
- Calls tools while the caller is still talking.
- Sub-600ms response time.
- Simple per-minute pricing with no token fees.
- 50+ languages with mid-call switching.
What could be better
- No free plan.
- Pro needs a $200 monthly minimum.
- Aimed at businesses, not consumers.
- Benchmarks are mostly Phonic's own.
Key features
- Phonic-1 model: Speech-to-speech for agentic tasks.
- Mid-speech tool calls: Answers are ready when the caller stops.
- Multilingual: 50+ languages, same voice.
- Builder options: API, SDKs, CLI and no-code builder.
- Compliance: SOC 2 Type II, HIPAA BAA, zero data retention.
- Deployment: Phonic Cloud to self-hosted.
Screenshots
Full review
Independent and reader-supported. Some links may earn us a commission; it never changes our scores. Editorial policy
Phonic AI builds speech-to-speech voice agents on its Phonic-1 model, launched October 29, 2026. Pricing is per minute: $0.07 on Starter and $0.065 on Pro.
Quick answer: Phonic is a startup that only builds speech-to-speech models for voice agents, such as phone support lines. Its new model, Phonic-1, hears the caller’s audio directly instead of working from a transcript, and it can call tools while the caller is still talking. Phonic says it took the #1 spot on the Artificial Analysis Speech to Speech Index. Alongside the launch, the company announced a $27.5M Series A led by First Round. We checked the prices on the Phonic pricing page on November 2, 2026.

What Phonic-1 does differently
Most voice agents chain three models: one transcribes, one decides what to say and one speaks. Phonic argues that a transcript loses the frustration in a caller’s voice or a long pause before an answer. Phonic-1 replaces the chain with one model. According to Phonic’s docs and launch post, it offers:
- Tool calls mid-sentence: it can look things up while the caller speaks, so the answer is ready when they stop.
- Low latency: under 600ms from speech in to speech out, even on long calls.
- 50+ languages: it can switch language mid-call and keep the same voice.
- Accurate alphanumerics: better pronunciation of URLs, emails and codes.
- Enterprise basics: SOC 2 Type II, HIPAA with a BAA, zero data retention and a 99.9% uptime SLA.
On Sierra’s multilingual full-duplex τ³-Voice benchmark, Phonic reports 83.2 for Phonic-1, against 79.9 for GPT-Live-1 and 68.5 for Gemini 3.8 Live. In third-party blind tests, listeners still preferred a real human about 60% of the time, which Phonic openly lists as work to do.
Phonic AI Pricing (2026)
| Plan | Price | Minimum | Includes |
|---|---|---|---|
| Starter | $0.07 a minute | None | 10 concurrent calls, 10 agents, 3 seats, unlimited voices |
| Pro | $0.065 a minute | $200 a month (about 3,100 minutes) | 50 concurrent calls, 50 agents, 30 seats, Slack Connect support |
| Enterprise | Custom | Volume discounts | Custom voices, on-prem or VPC, HIPAA BAAs, audit logs |

The per-minute rate covers everything: Phonic doesn’t add LLM pass-through or per-token charges. That makes cost easy to forecast. A 5-minute support call costs about $0.35 on Starter. For a sense of how other real-time voice stacks compare, read our ElevenLabs Realtime vs GPT Realtime breakdown.

Who Phonic is for
Phonic is a developer and business tool, not a consumer app. Teams build agents through the API, TypeScript and Python SDKs, a CLI or a no-code agent builder, using the model ID phonic-1. Phonic says customers include Ashby, Maven AGI and Rho across support, healthcare, recruiting and logistics. If you just want a voice for videos or narration, a text-to-speech tool from our best free AI voice generators list is a better fit, and Voxtral 3 covers Mistral’s speech option.
Limitations
- No free plan: Starter has no minimum, but every minute is billed.
- Built for businesses: there’s no consumer chat app.
- Self-reported results: most benchmarks come from Phonic’s launch post.
- Naturalness gap: humans still beat it in blind listening tests.
Phonic AI FAQ
Is Phonic AI free?
No. There’s no free plan, but Starter has no minimum, so you only pay $0.07 for each minute you use while building and testing.
•
How much does Phonic AI cost?
Starter is $0.07 a minute with no minimum. Pro is $0.065 a minute with a $200 monthly minimum. Enterprise pricing is custom.
•
What is Phonic-1?
Phonic-1 is Phonic’s speech-to-speech model for voice agents, launched October 29, 2026. It works on audio directly and can call tools while the caller is talking.
•
How fast is Phonic-1?
Phonic says Phonic-1 responds in under 600ms from speech in to speech out, even on long multi-step calls.
•
What languages does Phonic support?
Phonic-1 is trained on 50+ languages and can switch languages mid-conversation while keeping the same voice.