Browse tools

Developers working with Chinese models get a huge multimodal model at modest API prices; whether UK sign-up works wasn't checked.

Kenning's take

What it does

ERNIE 5.0 is a 2.4 trillion-parameter model from Baidu trained from scratch on text, images, video and audio together, so one model handles understanding and generation across them. Only about 3% of its parameters are active for each token. Baidu unveiled it in November 2025, with a preview in the ERNIE Bot app and on its Qianfan cloud API.

  • 2.4 trillion parameters, mixture-of-experts
  • Text, image, video and audio in one model
  • 128k token context
  • Available through Baidu Qianfan API

Who it suits. Developers and businesses using Chinese AI models, and ERNIE Bot users.

Key facts

Category
AI models
Made by
Baidu
Based in
China
Released
13 November 2025
How to get it
Preview in ERNIE Bot for the public; API on Baidu AI Cloud Qianfan (model id ernie-5.0)
Works on
Web, API
API price
$0.84 per million input tokens, $3.38 per million output (input up to 32k); vision input $1.40, output $5.60 per million

Reviews and write-ups

FAQ

When was ERNIE 5.0 released?

Baidu unveiled ERNIE 5.0 on 13 November 2025 at Baidu World 2025.

How can ERNIE 5.0 be used?

A preview is in ERNIE Bot, and developers can call it through the Baidu Qianfan API.

How big is ERNIE 5.0?

Baidu says it has 2.4 trillion parameters, with about 3% active per token.

Facts from the maker's own pages and the sources listed. Checked 2026-10-11.

Similar tools

ACE-Step 1.5

Open-source music generation model that writes full songs with lyrics and runs on ordinary consumer GPUs.

AlphaGenome Atlas

Google DeepMind database of AI-predicted effects for every possible single-letter change in human DNA.

Amazon Nova 2

Amazon's second family of Nova AI models, covering reasoning, speech conversation and combined text-and-image generation.

Cisco Antares

Cisco's small open-weight models that find where known security vulnerabilities sit in a codebase, running locally.