Baidu: ERNIE 4.5 300B A47B

Text input Text output
Author's Description

ERNIE-4.5-300B-A47B is a 300B parameter Mixture-of-Experts (MoE) language model developed by Baidu as part of the ERNIE 4.5 series. It activates 47B parameters per token and supports text generation in both English and Chinese. Optimized for high-throughput inference and efficient scaling, it uses a heterogeneous MoE structure with advanced routing and quantization strategies, including FP8 and 2-bit formats. This version is fine-tuned for language-only tasks and supports reasoning, tool parameters, and extended context lengths up to 131k tokens. Suitable for general-purpose LLM applications with high reasoning and throughput demands.

Key Specifications
Cost
$$$
Context
123K
Parameters
300B
Released
Jun 30, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Stop Presence Penalty Logit Bias Top P Temperature Seed Min P Frequency Penalty Max Tokens
Performance Summary

Baidu's ERNIE-4.5-300B-A47B demonstrates moderate speed performance, ranking in the 27th percentile across various benchmarks. However, it offers competitive pricing, typically providing cost-effective solutions and ranking in the 70th percentile. The model exhibits a notable strength in Coding, achieving 90.0% accuracy, placing it in the 81st percentile for this category. Its instruction following capabilities are also strong, with 68.0% accuracy, ranking in the 83rd percentile. These results suggest the model is well-suited for tasks requiring precise adherence to instructions and programming knowledge. Conversely, the model shows significant weaknesses in several areas. Its performance in Email Classification is very low at 29.0% accuracy (6th percentile), indicating a struggle with nuanced categorization. Similarly, Reasoning (16.0% accuracy, 11th percentile), Ethics (33.0% accuracy, 14th percentile), and General Knowledge (37.5% accuracy, 16th percentile) benchmarks reveal considerable limitations in complex problem-solving, ethical judgment, and broad factual recall. While its cost-effectiveness is a clear advantage, the model's inconsistent accuracy across different domains suggests it is best utilized for its strong coding and instruction-following capabilities rather than general-purpose reasoning or classification tasks.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.28
Completion $1.1

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Novita
Novita | baidu/ernie-4.5-300b-a47b 123K $0.28 / 1M tokens $1.1 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Free Executions Accuracy Cost Duration
Other Models by baidu