AssemblyAI Pricing 2026: $0.21/hr Plans Explained

A common search query for developers is: "AssemblyAI pricing per minute".

On the surface, it looks simple. But once you start adding features like Speaker Diarization, PII Redaction, and Sentiment Analysis, the bill gets complicated.

Here is the definitive guide to AssemblyAI pricing in 2026.

The Base Rate: Transcription (Updated September 2026)

AssemblyAI bills pre-recorded audio per audio hour and streaming per WebSocket session (idle connection time counts — close sockets promptly). Current flagship rates:

• Async (Universal-3.5 Pro): $0.21 / hour flat (~$0.0035/min). The older Universal-2 is $0.15/hr.
• Real-Time (Universal-3.5 Pro Realtime): $0.45 / hour (~$0.0075/min). Budget Universal-Streaming models are $0.15/hr.
• Voice Agent API (STT+LLM+TTS pipeline): $4.50 / hour flat.

Comparison (async):

• Deepgram Nova-3: $0.0043–0.0048 / min ($0.26–0.29/hr) — slightly more than AssemblyAI's flat rate at list price.
• Google Cloud: $0.016 / min (4x more expensive).
• OpenAI Whisper API: $0.006 / min.

Verdict: The old "$0.37/hr premium" is retired — no current tier prices there. At list price, AssemblyAI's flat $0.21/hr now undercuts Nova-3 pay-as-you-go for async work. Deepgram fights back with promo streaming rates and ~20% Growth-plan discounts.

The "Hidden" Costs: Audio Intelligence

This is where AssemblyAI makes its money. Unlike Deepgram (where many features are bundled), AssemblyAI charges extra for "Intelligence" models.

These are add-ons that stack on top of the base rate (per audio hour, async):

1. PII Text Redaction: +$0.08 / hr
2. Sentiment Analysis: +$0.02 / hr
3. Entity Detection: +$0.08 / hr
4. Speaker Diarization: +$0.02 / hr (standard)
5. Medical Mode: +$0.15 / hr
6. Summarization / Auto Chapters: +$0.03–0.08 / hr — Universal-2 only and deprecated; use the LLM Gateway instead.

The "Stacking" Effect

Let's say you are building a Call Center Analytics tool on Universal-3.5 Pro. You need:

1. Transcription ($0.21)
2. Redaction (to hide credit card numbers) ($0.08)
3. Sentiment (to detect angry customers) ($0.02)
4. Diarization (who said what) ($0.02)

Total Cost: $0.33 / hour (~$0.0055/min).

The stack adds ~57% to the base — far less dramatic than the old per-minute add-on era, and still under a single hour of Google Cloud transcription.

Volume Discounts (The "Enterprise" Tier)

Like all API providers, the list price is for suckers (or startups).

Once you exceed 10,000 hours per month, you enter the negotiation zone.

• Target Price: You should aim to get the base transcription rate down to $0.003 - $0.004 / min.
• Bundling: Try to negotiate the Audio Intelligence features into a flat fee or a reduced bundle rate.

Is It Worth It?

Yes, if:

• You need state-of-the-art PII Redaction. AssemblyAI's redaction is widely considered better than Deepgram's regex-heavy approach.
• You need Speaker Diarization on messy audio. AssemblyAI's diarization (splitting Speaker A vs Speaker B) handles interruptions better than most open-source models.
• You want a "batteries included" NLP pipeline without managing your own LLM for summaries.

No, if:

• You just need cheap batch text and live on promo rates. Deepgram Nova-3 with a Growth commit can beat $0.21/hr at volume.
• You need sub-200ms turn-taking specifically — benchmark Flux-style endpointing against Universal-3.5 Pro Realtime on your traffic.

Summary Table (September 2026)

Feature	Price
Async (Universal-3.5 Pro)	$0.21 / hr (~$0.0035/min)
Async (Universal-2)	$0.15 / hr
Streaming (3.5 Pro Realtime)	$0.45 / hr (~$0.0075/min)
Voice Agent API (all-in)	$4.50 / hr
+ Diarization (async)	+$0.02 / hr
+ PII Redaction (text)	+$0.08 / hr
+ Sentiment / Entity	+$0.02 / +$0.08 / hr
+ Medical Mode	+$0.15 / hr

Rates from AssemblyAI's pricing page, September 2026. Free tier: $50 credit, no card. Verify before budgeting — add-ons and promos move.

Related comparisons & reviews

• Deepgram vs. AssemblyAI (2026): Which STT API Should You Choose?
• AssemblyAI Universal-2 Review: Is Accuracy Worth the Extra Latency?
• Best AI Scribes of 2026: The Ultimate Comparison Guide