ERNIE 5.0 Thinking Preview

Name: ERNIE 5.0 Thinking Preview
Price: 6 CNY
Availability: InStock
Author: Baidu

PreviewReasoning modelERNIE 5.0

Release date: 2025-11-01Updated: 2026-07-17Views: 517

Live demoGitHubHugging FaceCompare

Parameters

2.4T

Context length

128K

Multilingual

Supported

Reasoning ability

2/5

ERNIE 5.0 Thinking Preview is a reasoning model from Baidu, released in 2025-11. It accepts text, image, audio, and video input and returns text output. The cataloged parameter count is 2.4T. The recorded context window is 128K, and the recorded maximum output is 64K. Cataloged capabilities include Reasoning model and Multilingual. The model weights are proprietary and are not published for download. The model is listed as a preview, so access and production terms may remain limited. Use the linked references to confirm current access, licensing, and provider-specific limits.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

ERNIE 5.0 Thinking Preview

Model basics

Reasoning traces

Supported

Thinking modes

Thinking Mode (Default)

Context length

128K tokens

Max output length

64K tokens

Model type

Reasoning model

Modality (in / out)

Text, Image, Audio, Video → Text

Release date

2025-11-01

Model file size

No data

MoE architecture

Yes

Total params / Active params

2.4T / Not applicable

Knowledge cutoff

No data

ERNIE 5.0 Thinking Preview

Open source & experience

Code license

Proprietary

Weights license

Proprietary

GitHub repo

N/A

Hugging Face

N/A

Live demo

ernie.baidu.com

ERNIE 5.0 Thinking Preview

Official resources

Paper

DataLearnerAI blog

N/A

ERNIE 5.0 Thinking Preview

API details

API speed

3/5

💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.

Learn about pricing modes

Standard

Type	Condition	Input	Output
Text	Context <= 32000	¥6.00/ 1M tokens	¥24.00/ 1M tokens
Text	Context > 32000, Context <= 128000	¥10.00/ 1M tokens	¥40.00/ 1M tokens

“—” means the modality is not billed in that direction, or the vendor has not published a price for it.

ERNIE 5.0 Thinking Preview

Benchmark Results

ERNIE 5.0 Thinking Preview currently shows benchmark results led by LiveCodeBench (45 / 250, score 81.20), AIME2025 (76 / 215, score 85), τ²-Bench - Telecom (98 / 264, score 83.90). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

General Knowledge

2 evaluations

Benchmark / mode

Score

Rank/total

HLE

Thinking Mode

13.30

355 / 563

CritPt

Thinking Mode

1.40

135 / 200

General Evaluation

1 evaluations

Benchmark / mode

Score

Rank/total

GPQA Diamond

Thinking Mode

77.70

259 / 462

Coding and Software Engineer

1 evaluations

Benchmark / mode

Score

Rank/total

LiveCodeBench

Thinking Mode

81.20

45 / 250

Math and Reasoning

1 evaluations

Benchmark / mode

Score

Rank/total

AIME2025

Thinking Mode

76 / 215

Agent Level Benchmark

2 evaluations

Benchmark / mode

Score

Rank/total

τ²-Bench - Telecom

Thinking ModeTools

83.90

98 / 264

Terminal Bench Hard

Thinking ModeTools

128 / 244

Instruction Following

1 evaluations

Benchmark / mode

Score

Rank/total

IF Bench

Thinking Mode

41.40

208 / 282

Claw-style Agent Evaluation

1 evaluations

Benchmark / mode

Score

Rank/total

Claw Bench

Thinking Mode

29 / 29

Multimodal Understanding

1 evaluations

Benchmark / mode

Score

Rank/total

MMMU-Pro

Thinking Mode

64.60

155 / 227

View benchmark analysis Compare with other models

ERNIE 5.0 Thinking Preview

Publisher

Baidu

View publisher details

ERNIE 5.0 Thinking Preview

Model Overview

ERNIE 5.0 Thinking Preview is a reasoning model from Baidu, released in 2025-11.

It accepts text, image, audio, and video input and produces text output. Its cataloged capabilities include Reasoning model and Multilingual. The cataloged parameter count is 2.4T. The recorded context window is 128K, and the recorded maximum output is 64K.

The model is listed as a preview, so access and production terms may remain limited. The model weights are proprietary and are not published for download. The page records 4 API pricing rules from the listed provider; current provider pricing and conditions should be checked before deployment. The evaluation section contains 1 cataloged benchmark result with its recorded mode and score. The page links 2 release, model-card, repository, or provider references for checking the underlying claims. Specifications, availability, and prices can change; undisclosed values are intentionally left unstated.

ERNIE 5.0 Thinking Preview

FAQ

What is ERNIE 5.0 Thinking Preview?

What input and output modalities does ERNIE 5.0 Thinking Preview support?

The current model record lists text, image, audio, and video as input and text as output.

What are the main recorded specifications for ERNIE 5.0 Thinking Preview?

The cataloged parameter count is 2.4T. The recorded context window is 128K, and the recorded maximum output is 64K. Fields without a source-backed value remain undisclosed.

Does ERNIE 5.0 Thinking Preview have API pricing?

The page records 4 API pricing rules from the listed provider; current provider pricing and conditions should be checked before deployment.

Are benchmark results available for ERNIE 5.0 Thinking Preview?

The evaluation section contains 1 cataloged benchmark result with its recorded mode and score. Compare only results that use the same benchmark version and evaluation mode.

Is ERNIE 5.0 Thinking Preview open source?

The model weights are proprietary and are not published for download. Review the linked license text before commercial or derivative use.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.