ERNIE 4.5
ERNIE 4.5 is Baidu's large language model, offering strong performance in understanding, generation, logical reasoning, and coding. The model supports text and multimodal inputs. It demonstrates competitive results on major benchmarks and is optimized for both Chinese and English language tasks.
Key Specifications
Parameters
21.0B
Context
-
Release Date
June 24, 2025
Average Score
60.3%
Timeline
Key dates in the model's history
Announcement
June 24, 2025
Last Update
January 29, 2026
Today
March 25, 2026
Technical Specifications
Parameters
21.0B
Training Tokens
-
Knowledge Cutoff
-
Family
-
Capabilities
MultimodalZeroEval
Benchmark Results
Model performance metrics across various tests and benchmarks
General Knowledge
Tests on general knowledge and understanding
Winogrande
Accuracy • Self-reported
Reasoning
Logical reasoning and analysis
GPQA
Accuracy • Self-reported
Other Tests
Specialized benchmarks
ARC-E
Accuracy • Self-reported
PIQA
Accuracy • Self-reported
License & Metadata
License
proprietary
Announcement Date
June 24, 2025
Last Updated
January 29, 2026
Similar Models
All Models
Llama-3.3 Nemotron Super 49B v1
NVIDIA
49.9B
Best score:0.7 (GPQA)
Released:Mar 2025
Phi-3.5-MoE-instruct
Microsoft
60.0B
Best score:0.9 (ARC)
Released:Aug 2024
LongCat-Flash-Lite
Meituan
68.5B
Best score:0.9 (MMLU)
Released:Feb 2026
Phi 4 Reasoning
Microsoft
14.0B
Best score:0.9 (HumanEval)
Released:Apr 2025
Hermes 3 70B
Nous Research
70.0B
Best score:0.8 (MMLU)
Released:Aug 2024
Nemotron 3 Nano (30B A3B)
NVIDIA
32.0B
Best score:0.8 (GPQA)
Released:Dec 2025
Price:$0.06/1M tokens
Llama 3.1 Nemotron 70B Instruct
NVIDIA
70.0B
Best score:0.8 (MMLU)
Released:Oct 2024
Phi 4
Microsoft
14.7B
Best score:0.8 (MMLU)
Released:Dec 2024
Price:$0.07/1M tokens
Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.