Script Planner Assistant▸ 
AI Apps 
AI News ERNIE 4.5 vs ChatGPT 4.5: A Comparative Analysis
小编 ·
ERNIE 4.5 vs ChatGPT 4.5: A Comparative Analysis
On March 16, 2025, Baidu released Ernie 4.5, which can understand text, images, audio, video, and other types of content, with comprehensively upgraded language capabilities. Prior to this, Openai launched GPT-4.5 on February 28, 2025. What are the features and differences between these two models?
Features of Ernie 4.5 and chatGPT 4.5
Ernie 4.5
Ernie 4.5 is Baidu's first natively multimodal large model, featuring outstanding multimodal understanding capabilities and comprehensively upgraded language abilities, including:
Strong multimodal capabilities: Can process text, images, audio, video, and other types of data, making it suitable for creative writing, video script generation, and other multimodal tasks.
High cost-effectiveness: Low API pricing at 0.004 RMB per 1,000 input tokens and 0.016 RMB per 1,000 output tokens.
Excellent logical reasoning: Performs exceptionally well in complex logical reasoning and long chain-of-thought tasks.
chatGPT 4.5
chatGPT 4.5 is the latest version released by OpenAI, featuring powerful language understanding and generation capabilities, widely applied to various natural language processing tasks.
Excellent text generation: Excels at high-quality text generation, suitable for creative writing, copywriting, and more.
Improved emotional intelligence: Enhanced emotional understanding and interaction capabilities, making it more "human-like."
Limited performance improvement: Training compute has increased significantly, but cost-effectiveness is low and the price is high.
Comparative Analysis of Ernie 4.5 and GPT-4.5
| Comparison Dimension | Ernie 4.5 | chatGPT 4.5 |
|---|---|---|
| Multimodal Capability | A natively multimodal large model that supports graphical reasoning, chart analysis, and understanding of memes, comics, songs, movies, etc. | Primarily focused on text generation and understanding, with relatively weak multimodal capabilities. |
| Text and Logical Reasoning Capabilities | Outperforms GPT-4.5 on multiple benchmarks, with an average score of 79.6, demonstrating strong logical reasoning abilities. | Excels in text generation, but has weaker logical reasoning capabilities, particularly in complex reasoning tasks. |
| Performance Optimization | Achieves end-to-end optimal tuning through joint optimization with PaddlePaddle and Ernie. | Training compute is 10 times that of GPT-4o, but performance improvement is limited. |
| Price | Affordable API pricing at 0.004 RMB per 1,000 input tokens and 0.016 RMB per 1,000 output tokens. | Relatively expensive, at $75 per million input tokens and $150 per million output tokens. |
| Application Scenarios | Suitable for multimodal understanding and complex logical reasoning tasks, such as creative writing, data analysis, and more. | Suitable for text generation, creative writing, sentiment analysis, and other tasks, but with limited performance on multimodal and complex reasoning tasks. |
Overall, Ernie 4.5 is well-suited for multimodal tasks and complex logical reasoning, offering high cost-effectiveness. GPT-4.5 is suited for high-quality text generation and emotional interaction, but comes at a higher cost.
