Baidu ERNIE vs Deepgram: Token Pricing, Speed & Intelligence
Full comparison of Baidu ERNIE and Deepgram — live token pricing, latency, throughput, context window, strengths, weaknesses, and best use cases. Rates sourced from public APIs — verify before use.
Baidu ERNIE
ERNIE 5 — Baidu's frontier Chinese-English bilingual LLM API
Baidu ERNIE is the LLM API from Baidu, China's largest search company. The ERNIE 5 series includes ERNIE-5.1 (balanced) and ERNIE-5.0 (frontier), both optimised for Chinese-English bilingual tasks, knowledge retrieval, and enterprise applications. Pricing is in USD via the international API.
Deepgram
Nova-3 speech-to-text API — real-time and batch transcription at low latency
Deepgram is a speech-to-text API provider specialising in low-latency audio transcription. Its Nova-3 model delivers real-time streaming transcription with high accuracy across accents and domains. Nova-3 Medical is tuned for clinical vocabulary. Deepgram also hosts Whisper Large for teams that need an open-weights baseline. Pricing is per audio hour rather than per token.
Key metric comparison
—
—
—
—
—
—
—
—
—
—
—
—
Live token pricing
Strengths and weaknesses
Baidu ERNIE
Deepgram
Key differentiators
ERNIE 5 is the leading Chinese-English bilingual LLM API, backed by Baidu's search knowledge graph — the go-to choice for applications targeting Chinese-language users.
Purpose-built speech-to-text API with Nova-3 at ~300 ms latency and a medical-domain variant — priced per audio hour rather than per token.
Frequently asked questions
Baidu ERNIE FAQs
What is ERNIE?
ERNIE (Enhanced Representation through Knowledge Integration) is Baidu's LLM series. ERNIE 5 is the current generation, with ERNIE-5.1 (balanced) and ERNIE-5.0 (frontier) available via the Wenxin Qianfan API.
How much does the ERNIE API cost?
ERNIE-5.1 is $0.56/$2.53 per 1M input/output tokens. ERNIE-5.0 is $1.40/$5.60 per 1M tokens. See the Baidu Cloud pricing page for current rates.
Is ERNIE good for English tasks?
ERNIE 5 handles English competently but is primarily optimised for Chinese. For English-only workloads, frontier Western models typically outperform ERNIE. For Chinese or bilingual Chinese-English tasks, ERNIE is among the strongest options available.
Deepgram FAQs
How is Deepgram priced?
Deepgram charges per audio hour rather than per token. Nova-3 is $0.218/hour, Nova-3 Medical is $0.435/hour, and Whisper Large is $0.218/hour. See deepgram.com/pricing for current rates.
What is Deepgram Nova-3?
Nova-3 is Deepgram's latest speech recognition model, offering high accuracy with approximately 300 ms P50 latency for real-time streaming. It supports 30+ languages and is optimised for conversational and call-centre audio.
Does Deepgram support real-time transcription?
Yes. Deepgram's streaming API delivers transcription results with ~300 ms latency, making it suitable for live captions, voice agents, and real-time call analytics.
Provider resources
Baidu ERNIE — ERNIE 5 — Baidu's frontier Chinese-English bilingual LLM API
Baidu ERNIE is the LLM API from Baidu, China's largest search company. The ERNIE 5 series includes ERNIE-5.1 (balanced) and ERNIE-5.0 (frontier), both optimised for Chinese-English bilingual tasks, knowledge retrieval, and enterprise applications. Pricing is in USD via the international API.
ERNIE 5 is the leading Chinese-English bilingual LLM API, backed by Baidu's search knowledge graph — the go-to choice for applications targeting Chinese-language users.
Deepgram — Nova-3 speech-to-text API — real-time and batch transcription at low latency
Deepgram is a speech-to-text API provider specialising in low-latency audio transcription. Its Nova-3 model delivers real-time streaming transcription with high accuracy across accents and domains. Nova-3 Medical is tuned for clinical vocabulary. Deepgram also hosts Whisper Large for teams that need an open-weights baseline. Pricing is per audio hour rather than per token.
Purpose-built speech-to-text API with Nova-3 at ~300 ms latency and a medical-domain variant — priced per audio hour rather than per token.
Key strengths compared
Baidu ERNIE
- Strong Chinese-English bilingual performance — best-in-class for Chinese language tasks
- Deep integration with Baidu Search knowledge graph
- ERNIE-5.1 at $0.56/$2.53 per 1M tokens is competitive for a frontier-tier bilingual model
- Primarily optimised for Chinese — English performance trails frontier Western models
- Limited ecosystem outside China; fewer third-party integrations
Deepgram
- Nova-3 delivers industry-leading accuracy with ~300 ms P50 latency for real-time streaming
- Nova-3 Medical tuned for clinical and healthcare vocabulary
- Per-audio-hour pricing is predictable for high-volume transcription workloads
- Speech-to-text only — no LLM chat, image, or embedding models
- No per-token pricing; cost structure differs from general-purpose LLM APIs
Key weaknesses compared
Baidu ERNIE
- Primarily optimised for Chinese — English performance trails frontier Western models
- Limited ecosystem outside China; fewer third-party integrations
- Documentation and support primarily in Chinese
Deepgram
- Speech-to-text only — no LLM chat, image, or embedding models
- No per-token pricing; cost structure differs from general-purpose LLM APIs
- Smaller ecosystem than general-purpose providers
Provider category context
Baidu ERNIE is a frontier lab, founded in 2000. Deepgram is a frontier lab, founded in 2015. Both are frontier lab providers — the comparison is primarily about pricing, model selection, and feature differentiation within the same tier.
How to choose between them
Both Baidu ERNIE and Deepgram are frontier labs with proprietary models. Choose based on benchmark performance for your specific task: Baidu ERNIE leads on strong chinese-english bilingual performance — best-in-class for chinese language tasks, while Deepgram leads on nova-3 delivers industry-leading accuracy with ~300 ms p50 latency for real-time streaming. For cost-sensitive workloads, compare the cheapest model tier from each provider in the pricing table above — the gap between efficient-tier models is often larger than between flagship models.