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
Supported Parameters
This model supports the following parameters:
Performance Summary
Baidu's ERNIE 4.5 300B A47B demonstrates moderate speed performance, ranking in the 21st percentile across benchmarks, indicating it is not among the fastest models available. However, it offers competitive pricing, placing in the 55th percentile. The model exhibits a notable strength in hallucination avoidance, achieving 96.0% accuracy, suggesting a strong ability to acknowledge uncertainty. It also performs well in Instruction Following (68.0% accuracy, 78th percentile) and Coding (90.0% accuracy, 73rd percentile), indicating proficiency in structured tasks and programming knowledge. Conversely, the model shows significant weaknesses in several areas. Its performance in General Knowledge (37.5% accuracy, 15th percentile), Ethics (33.0% accuracy, 13th percentile), Email Classification (29.0% accuracy, 6th percentile), and Reasoning (34.0% accuracy, 22nd percentile) is considerably low. The extremely low accuracy in Email Classification is particularly concerning. While its Mathematics accuracy is moderate (80.4%), the duration for this benchmark is very high, placing it in the 9th percentile for speed. Overall, ERNIE 4.5 300B A47B appears well-suited for tasks requiring strong instruction adherence and coding capabilities, but struggles with broader knowledge, ethical reasoning, and 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 |
SiliconFlow
|
SiliconFlow | baidu/ernie-4.5-300b-a47b | 131K | $0.28 / 1M tokens | $1.1 / 1M tokens |
Benchmark Results
Benchmark | Category | Reasoning | Strategy | Free | Executions | Accuracy | Cost | Duration |
---|
Other Models by baidu
|
Released | Params | Context |
|
Speed | Ability | Cost |
---|---|---|---|---|---|---|---|
Baidu: ERNIE 4.5 21B A3B | Aug 12, 2025 | 21B | 120K |
Text input
Text output
|
★★★★ | ★★★ | $$ |
Baidu: ERNIE 4.5 VL 28B A3B | Aug 12, 2025 | 28B | 30K |
Text input
Image input
Text output
|
★★★ | ★★★ | $$$ |
Baidu: ERNIE 4.5 VL 424B A47B | Jun 30, 2025 | 424B | 123K |
Text input
Image input
Text output
|
★★ | ★★★★ | $$$$ |