| Deepgram | models[5] | — → {"id":"flux-tts","name":"Flux TTS","scene":"tts","summary":"Conversation-native text to speech for real-time voice agents.","pricing":{"model_id":"flux-tts","billing":"per_unit_tiered","unit":"thousand_chars","currency":"USD","tier_mode":"whole","tiers":[{"upto":null,"rate":0.045}],"notes":"Pay As You Go and Growth are both $0.0450 per 1K characters. Deepgram offers extra credits on Flux TTS spend until 2026-12-31; not modelled.","source_url":null},"released":null} | source |
| Deepgram | models[1].pricing.notes | Nova-3 Monolingual streaming, Pay As You Go: $0.0048/min ($0.29/hr). promo: limited-time promotional rate, struck-through list price $0.0077/min; no end date published. Growth plan $0.0042/min (not modelled). → Nova-3 Monolingual streaming, Pay As You Go: $0.0077/min ($0.46/hr), checked against the live page on 2026-10-30. Growth plan $0.0065/min (not modelled). | source |
| Deepgram | models[1].pricing.addons.multilingual.note | Nova-3 Multilingual streaming is $0.0058/min versus $0.0048/min monolingual. → Nova-3 Multilingual streaming is $0.0092/min versus $0.0077/min monolingual. | source |
| Deepgram | models[1].pricing.addons.multilingual.value | 0.001 → 0.0015 | source |
| Deepgram | models[1].pricing.tiers[0].rate | 0.0048 → 0.0077 | source |
| Deepgram | notes | Pay As You Go list prices, USD. The default speech-to-text figure is Nova-3 pre-recorded (batch) monolingual, where speaker diarization is included; streaming Nova-3 is a separate model entry and its rates are temporarily discounted (Deepgram labels them limited-time rates, struck-through list price $0.0077/min monolingual, with no published end date). Growth plan (pre-paid credits from about $4K/year, roughly 12-15% cheaper) is not modelled. Voice Agent API is billed per minute of websocket connection time at the Standard tier; Advanced tier and bring-your-own LLM/TTS variants cost differently. Concurrency on Pay As You Go: 45 for TTS and Voice Agent, 150 for streaming STT. → Pay As You Go list prices, USD. The default speech-to-text figure is Nova-3 pre-recorded (batch) monolingual, where speaker diarization is included; streaming Nova-3 is a separate model entry with a higher per-minute rate. Growth plan (pre-paid credits from about $4K/year, roughly 12-15% cheaper) is not modelled. Voice Agent API is billed per minute of websocket connection time at the Standard tier; Advanced tier and bring-your-own LLM/TTS variants cost differently. Concurrency on Pay As You Go: 45 for TTS and Voice Agent, 150 for streaming STT. | source |