Model Comparator

Compare up to four models using the same criteria.

ERNIE 5.0 Thinking Preview
Baidu
DeepSeek V3.2 Speciale
DeepSeek

Differences only

✓ Best published value among measured models. — Data unavailable. These metrics do not establish an overall ranking.

Model Comparator
Metrics	ERNIE 5.0 Thinking Preview Baidu	DeepSeek V3.2 Speciale DeepSeek	Add a model · 2/4
Intelligence	14.3	Best: 14.5
Math	85	Best: 96.7
Context window	128K	128K
Reasoning	Yes	Yes
Open weights	No	Yes

Intelligence

ERNIE 5.0 Thinking Preview: 14.3
DeepSeek V3.2 Speciale: Best: 14.5

Math

ERNIE 5.0 Thinking Preview: 85
DeepSeek V3.2 Speciale: Best: 96.7

Context window

ERNIE 5.0 Thinking Preview: 128K
DeepSeek V3.2 Speciale: 128K

Reasoning

ERNIE 5.0 Thinking Preview: Yes
DeepSeek V3.2 Speciale: Yes

Open weights

ERNIE 5.0 Thinking Preview: No
DeepSeek V3.2 Speciale: Yes

Each reasoning level remains a separate configuration.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.