Assemblyai

New FAQ entry added asking what's included in the Voice Agent API's flat $4.50/hr rate. The question is newly visible in OCR; the detailed 'included features' answer text is markdown-only and not painted on screen.

Pre-recorded model 'Universal-3 Pro' renamed to 'Universal-3.5 Pro'; description now highlights 18-language support, native code switching, and improved speaker diarization. Price unchanged at $0.21/hr.

New realtime STT model announced: a site-wide banner on the pricing page now reads "NEW - Universal-3.5 Pro Realtime is here." with a Learn more link; no such banner appeared in the prior version.

Free plan pre-recorded audio limit increased from 185h to 333h; separate pre-recorded and streaming limits consolidated into one unified 333h allowance for pre-recorded or streaming audio.

Medical Mode add-on introduced for speech-to-text models; optimizes transcription for medical terminology in healthcare conversations.

Pre-recorded STT models restructured: 'Universal' renamed to 'Universal-2' (price unchanged at $0.15/hr); 'Slam-1 Beta' renamed to 'Universal-3 Pro Default' with price reduced from $0.27/hr to $0.21/hr. Universal-3 Pro is now the flagship model. New Whisper-Streaming model added at $0.30/hr. Supports 99+ languages. Includes Keyterms Prompting add-on ($0.04/hr) and Speaker Diarization add-on ($0.12/hr).

Quarterly roll-up · includes 5 weeklies

Added Claude 4.6 Sonnet model to LLM Gateway at $3.00/1m tokens (input), $15.00/1m tokens (output); Added Claude 4.6 Opus model to LLM Gateway at $5.00/1m tokens (input), $25.00/1m tokens (output); Added Claude 4.5 Opus model to LLM Gateway at $5.00/1m tokens (input), $25.00/1m tokens (output)

Introduced Whisper-Streaming model at $0.30/hr as open-source Whisper enhanced with AssemblyAI infrastructure supporting 99+ languages Introduced Universal-3 Pro Streaming model at $0.45/hr for voice agents with best-in-class accuracy and advanced prompting capabilities

Pay as you go plan now highlights 'Speech Understanding and LLM Gateway' instead of 'Audio Intelligence and LeMUR', reflecting updated product naming.

Added Gemini 2.5 Flash Lite model to LLM Gateway at $0.10/1M input tokens, $0.40/1M output tokens

Guardrails introduced as new product category - consolidates safety features: Profanity Filtering, PII Audio/Text Redaction, Content Moderation LLM Gateway expanded with new models: GPT-5.2, GPT-5.1, GPT-5, GPT-5-Mini, GPT-5 Nano, GPT 4.1, gpt-oss-20b, gpt-oss-120b, ChatGPT-4o, Gemini 3 Pro, Gemini 3 Flash, Gemini 2.5 Flash, Gemini 2.5 Pro, Claude 4.5 Sonnet, Claude 4.5 Haiku

Free tier pre-recorded audio hours decreased from 416 hours to 185 hours Pre-recorded Speech-to-Text restructured from Best/Nano tiers to Slam-1/Universal models, both now priced at $0.27/hr (Best decreased from $0.37, Nano increased from $0.12)

Self-hosted deployments (On-prem, VPC) announced as coming soon for Custom plan Custom plan rate limits changed from 'Higher rate limits' to 'Customize rate limits - scale to any workload'

Claude Instant LLM model removed (was sunsetting on 10/28/24) Dual Channel feature renamed to Multichannel

Free tier offering changed from 'Get started at no cost' to 'Get started with $50 in free credits' LeMUR pricing restructured from generic tiers (LeMUR Default, LeMUR Claude 2.1, LeMUR Basic) to specific Claude model versions (Claude 3.5 Sonnet, Claude 3 Opus, Claude 3 Haiku, Claude 3 Sonnet, Claude 2.1, Claude 2.0, Claude Instant) with explicit per-model pricing

Audio Intelligence models (Entity Detection, Topic Detection, Key Phrases, PII Redaction, Sentiment Analysis, Content Moderation, Auto Chapters, Summarization) now available for Free in the Free tier Added concurrency starting at 200 files for Pay-as-you-go plan

Entity Detection decreased from $0.15/hour to $0.08/hour (47% decrease) Sentiment Analysis decreased from $0.12/hour to $0.02/hour (83% decrease)

Volume discount program introduced: Save up to 75% with committed usage reservations