LLM Reference

ERNIE Speed Pro on Baidu Qianfan

ERNIE · Baidu AI

Serverless

Last refreshed 2026-05-09. Next refresh: weekly.

Why use ERNIE Speed Pro on Baidu Qianfan?

Baidu Qianfan offers ERNIE Speed Pro with pay-as-you-go pricing at $0.04/1M input tokens. Baidu Qianfan is Baidu Intelligent Cloud's foundation model and Agent development platform, providing hosted API access to ERNIE and Qianfan models for text generation, reasoning, multimodal understanding, embeddings, and agent workflows.

Input / 1M
$0.044
Output / 1M
$0.089
Cache
Not sourced
Batch
Not sourced

Setup recipe

Docs fallback
Install
Use the provider REST API or SDK
Auth
Create a provider API key
Call
model: ernie-speed-pro-128k
Model ID
ernie-speed-pro-128k

Request example

Curated snippets for this provider are not sourced yet. Use Baidu Qianfan documentation with model ID ernie-speed-pro-128k.

Gotchas

  • Use provider model ID "ernie-speed-pro-128k", not the LLMReference slug "ernie-speed-pro".

Pricing

TypePrice (per 1M)
Input tokens$0.04
Output tokens$0.09

Capabilities

No model capability flags are currently sourced.

About ERNIE Speed Pro

ERNIE Speed Pro is Baidu's high-throughput lightweight model, available via the Qianfan API. It offers 128K context with efficient inference at very low cost ($0.044/1M input tokens). Pro-tier successor to the legacy free ERNIE Speed model. API model ID: ernie-speed-pro-128k. Exact release date not publicly documented.

FAQ

What does ERNIE Speed Pro cost on Baidu Qianfan?

On Baidu Qianfan, ERNIE Speed Pro costs $0.044 per 1M input tokens and $0.089 per 1M output tokens.

What is the context window for ERNIE Speed Pro on Baidu Qianfan?

ERNIE Speed Pro supports a 128k token context window on Baidu Qianfan.

What API model ID do I use for ERNIE Speed Pro on Baidu Qianfan?

Use the model ID ernie-speed-pro-128k when calling Baidu Qianfan's API.

Who created ERNIE Speed Pro?

ERNIE Speed Pro was created by Baidu AI as part of the ERNIE model family.

Is ERNIE Speed Pro open source?

ERNIE Speed Pro is not open source; the seed data lists it as proprietary.

Get Started

Model Specs

Released2025-01-01
Context128k
ArchitectureDecoder Only