Skip to main content
Rase AI Intelligence Newsletter

Models & assistants Model or AI service

Baidu opens API access to ERNIE 5.5 Flash Preview

ERNIE 5.5 Flash Preview is a hosted language model built on Baidu’s earlier Flash Base checkpoint. Baidu lists it on the Qianfan platform. [1]

Model access
Preview API available through Baidu Qianfan at announcement. [1]
License
Baidu announced planned Apache 2.0 weights. This announcement does not establish their availability. [1]

What the model does

Baidu announced the post-trained preview on November 11, 2026. It describes a mixture-of-experts model with a context window of one million tokens. [1]

Why it matters

API access lets teams assess the post-trained model on representative work before changing an existing integration. It is useful to compare the preview with other versions using the same materials. [1]

Who can use it

Developers testing coding or longer text tasks can evaluate the preview through Qianfan. The earlier Flash Base is a separate checkpoint with its own release. [1]

Key dates

  1. Baidu announces ERNIE 5.5 Flash Preview API access [1]

Ways to use it

Test the preview on a long written brief and verify its output against the source requirements. Record the model version because preview endpoints can change. [1]

Model limitations

Baidu’s launch post promises a later release of preview weights under Apache 2.0. This profile can confirm API access at announcement and leaves the weight release unverified. [1]

Sources and verification

Catalog evidence checked How we verify records

  1. Primary source Introducing ERNIE 5.5 Flash Preview · ERNIE

Related profiles

Models & assistants

Hy4

Tencent (Hunyuan)

It is a language model family with documented preview and subsequent variants.

Medium confidenceEvidence checked

Models & assistants

Qwen4-Max

Alibaba (Qwen team, Alibaba Cloud)

It provides model-based assistance for text and related work tasks.

High confidenceEvidence checked

Models & assistants

Gemini 4 Hydrogen

Google DeepMind

It processes text and other supported inputs for tasks such as code assistance and planning.

High confidenceEvidence checked