Skip to main content

ERNIE 4.5 300B A47B

BaiduProprietary

Description

ERNIE 4.5 is an enhanced version of ERNIE 4.0 with improved reasoning capabilities, better instruction following, and stronger performance across multiple domains including mathematics, coding, and creative writing. Features advanced knowledge integration, multilingual support, and optimized thinking capabilities. This 21B parameter model uses A3B quantization for efficient inference.

Release Date
2025-06-30
Parameters
21.0B
Context Length
123K
Modalities
text

Capability Radar

32
general
27
coding
55
reasoning
48
scienceest.
0
agents
0
multimodal

Science uses a reasoning proxy when dedicated science benchmarks are unavailable.

Rankings

Domain#RankScoreSource
Code Ranking311
23.0
AA
General Ranking343
32.0
AA
Math Reasoning163
54.0
AA
Reasoning101
29.0
LS
Science194
48.0
AA

Benchmark Scores (LLM Stats)

Biology

GPQA74.0%SR

Factuality

SimpleQA1.8%SR

Finance

MMLU41.9%SR
MMLU-Pro16.0%SR

General

ARC-E60.7%SR
PIQA55.2%SR
MMLU-Redux43.2%SR
C-Eval40.7%SR
ARC-C40.6%SR
MBPP+40.2%SR
CMMLU39.8%SR
AGIEval28.5%SR

Language

Winogrande51.3%SR
CLUEWSC48.6%SR
BBH30.4%SR

Math

DROP28.6%SR
GSM8k25.2%SR
MATH12.4%SR

Reasoning

HellaSwag33.0%SR
HumanEval+25.0%SR

AA Evaluation Indices

Math Index
41.3
Intelligence Index
15.0
Coding Index
14.5
Math 500
0.9
Gpqa
0.8
Mmlu Pro
0.8
Aime
0.5
Livecodebench
0.5
Aime 25
0.4
Ifbench
0.4
Scicode
0.3
Terminalbench Hard
0.1
Hle
0.0
Lcr
0.0
Tau2
0.0

LLM Stats Category Scores

Biology
70
Chemistry
70
Physics
60
General
40
Language
40
Reasoning
40
Finance
30
Healthcare
30
Legal
30
Math
30
Factuality
0

Pricing

Input Price$0.28 / 1M tokens
Output Price$1.1 / 1M tokens
Blended Price (3:1)$0.485 / 1M tokens

Speed

Tokens/sec24.6 tokens/s
Time to First Token1.71s
Time to Answer1.71s

Available Providers

(LS internal units)
ProviderInput PriceOutput Price
Novita400K4.0M

External Sources