Baidu model product
0 with improved reasoning capabilities, better instruction following, and stronger performance across multiple domains including mathematics, coding, and creative writing.
Updated Aug 12, 2026. Default version: ERNIE 4.5
Technical details for the model's default version.
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Benchmark scores for ERNIE 4.5.
| CLUEWSC | 48.6% | 03 | 3 | 0.0% | C | |
| MBPP+ | 40.2% | 04 | 4 | 0.0% | C | |
| CMMLU | 39.8% | 06 | 6 | 0.0% | C | |
| ARC-E | 60.7% | 08 | 8 | 0.0% | C | |
| AGIEval | 28.5% | 10 | 10 | 0.0% | C | |
| HumanEval+ | 25.0% | 10 | 10 | 0.0% | C | |
| PIQA | 55.2% | 11 | 11 | 0.0% | C | |
| BBH | 30.4% | 12 | 12 | 0.0% | C | |
| C-Eval | 40.7% | 18 | 18 | 0.0% | C | |
| Winogrande | 51.3% | 22 | 22 | 0.0% | C | |
| HellaSwag | 33.0% | 27 | 27 | 0.0% | C | |
| DROP | 28.6% | 30 | 30 | 0.0% | C | |
| ARC-C | 40.6% | 34 | 34 | 0.0% | C | |
| SimpleQA | 1.8% | 46 | 46 | 0.0% | C | |
| GSM8k | 25.2% | 48 | 48 | 0.0% | C | |
| MMLU-Redux | 43.2% | 48 | 48 | 0.0% | C | |
| MATH | 12.4% | 71 | 71 | 0.0% | C | |
| MMLU | 41.9% | 100 | 100 | 0.0% | C | |
| GPQA | 74.0% | 113 | 234 | 51.9% | C | |
| MMLU-Pro | 16.0% | 128 | 129 | 0.8% | C |
Preference and agent-evaluation results for the default version.
The default version does not have a matching Arena result yet.
Provider-specific output speed and catalog latency for ERNIE 4.5. Runtime does not affect the capability score.
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Official vendor API pricing appears first, followed by individual provider offers.
The default version has no current input or output token prices.
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Available versions of this model. The score column identifies the version used in the overall ranking.
| ERNIE 4.5 | 0.0 | 21B | 128K | 65.5K | No | Proprietary |
Key information about ERNIE 4.5 and its available data.
0 with improved reasoning capabilities, better instruction following, and stronger performance across multiple domains including mathematics, coding, and creative writing. Features advanced knowledge integration, multilingual support, and optimized thinking capabilities.
This 21B parameter model uses A3B quantization for efficient inference.
Data as of 2026-08-11.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
Common questions about ERNIE 4.5.
ERNIE 4.5's default version was released on Jun 25, 2025.
No official standard PAYG price is currently available for ERNIE 4.5.
ERNIE 4.5 was created by Baidu.
The default version has a 128K token context window.
No. The default version is not marked as having publicly available weights.
No provider offering is currently linked to the default version.
Nearby ranked alternatives include Gemini 1.0 Pro, Llama 3.1 8B, Qwen2 7B.