ERNIE 4.5 VL 424B A47B Check detailed information and pricing for AI models
Context Length 123,000 tokens, baidu from provided
123,000
Context Tokens
$0.42
Prompt Price
$1.25
Output Price
9/16
Feature Support
Model 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 using a heterogeneous MoE architecture and modality-isolated routing to enable high-fidelity cross-modal reasoning, image understanding, and long-context generation (up to 131k tokens). Fine-tuned with techniques like SFT, DPO, UPO, and RLVR, this model supports both “thinking” and non-thinking inference modes. Designed for vision-language tasks in English and Chinese, it is optimized for efficient scaling and can operate under 4-bit/8-bit quantization.
Basic Information
Developer
baidu
Model Series
Other
Release Date
2025-06-30
Context Length
123,000 tokens
Max Completion Tokens
16,000 tokens
Variant
standard
Pricing Information
Prompt Tokens
$0.42 / 1M tokens
Completion Tokens
$1.25 / 1M tokens
Data Policy
Supported Features
Supported (9)
Image Input
Top K
Seed
Frequency Penalty
Presence Penalty
Repetition Penalty
Min P
Logit Bias
Reasoning
Unsupported (7)
Response Format
Tool Usage
Logprobs
Top Logprobs
Structured Outputs
Web Search Options
Top A
Actual Usage Statistics
#325
Out of 353 total models
9.24M
Total Tokens Last 30 Days
3.08M
Daily Average Usage
-
Weekly Usage Change