Baidu

ERNIE 4.5 VL 424B A47B

by Baidutext+image->text1 endpoint
Neura Intelligence Index
38.9/ 100
Rank
#226
Confidence
low
Benchmarks
2(5% coverage)

Overview

ERNIE-4.5-VL-424B-A47B is a multimodal Mixture-of-Experts (MoE) model from Baidu’s ERNIE 4.5 series, featuring 424B total parameters with 47B active per token. It is trained jointly on text and image data...

Capabilities

Text generation
Image understanding
Extended reasoning
Long context
Top-P sampling
Stop sequences
Deterministic seed

Modalities

Input
ImageText
Output
Text

Technical Specifications

Context Window
123K tokens
Max Output
16K tokens
Knowledge Cutoff
2025-03-31
Tokenizer
Other
Uptime (24h)
10000.0%
Parameters
21000000000B
Country
🇨🇳CN
Release Date
June 25, 20251y

Supported Parameters

frequency_penaltyinclude_reasoningmax_tokenspresence_penaltyreasoningrepetition_penaltyseedstoptemperaturetop_ktop_p

Benchmark Performance

general

BenchmarkScoreSource
SimpleQA
1.8
verified

reasoning

BenchmarkScoreSource
GPQA Diamond
74
verified

Performance Over Time

Pricing

Per 1M tokens
Input$0.42
Output$1.25
Added
June 30, 2025
Last synced 9/18/2026