[whatmodel.sh](../../README.md) / ERNIE 4.5 300B-A47B
# ERNIE 4.5 300B-A47B
Baidu · released 2025-06 · 300B params (47B active) · apache-2.0
**score 60.1** ±1.3 · #200 of 732 · from 18 benchmark results across 1 source · † no independently run results verified
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks ERNIE 4.5 300B-A47B among every model on that task or test, worst to best.
task vs the others rank counts tests
coding ······●····· #1 of 1 15% 3
chat & writing ······●····· #1 of 1* 2.8% 1
math ······●····· #1 of 1 34% 4
coding · #1 of 1 · 15% of its score
test result vs the others in its run difficulty counts why
HumanEval+ (maker) 92.1% ▁·▁▁▂▃▄▄█▅▄● beats 138/145 easy · avg 82% 6.5% sharp · maxed out · small run
LiveCodeBench (version unspecifie… 38.8% ▁···▁▁▁●▇▃▄▁ beats 7/38 hard · avg 27% 5.3% at its level · small run
MBPP+ (maker) 81.0% ▁··▁▁▃▄█▆▄●▁ beats 115/120 easy · avg 70% 3.2% small run
chat & writing · #1 of 1 · 2.8% of its score
test result vs the others in its run difficulty counts why
IFEval (aggregation unspecified) … 88.0% ▁▂▄▂▂▄▆●█▅▂▁ beats 95/145 easy · avg 78% 2.8% small run
math · #1 of 1 · 34% of its score
test result vs the others in its run difficulty counts why
AIME 2025 (maker) 35.1% ▁·▁▁▁▂●▆█▇▅▁ beats 22/133 hard · avg 18% 11% sharp · at its level · small run
AIME 2024 (maker) 54.8% ▁▁▁▂▁▃▁●███▃ beats 25/85 hard · avg 29% 10% sharp · at its level · small run
GSM8K (maker) 96.6% ▂▂▂▂▃▄▅▇██▆● beats 205/206 very easy · avg 88% 8.4% sharp · maxed out · small run
MATH-500 (maker) 96.4% ▁▃▂▃▄▇▅▇●▄▂▁ beats 62/85 very easy · avg 87% 4.0% maxed out · small run
other tests (10)
test result vs the others in its run difficulty counts why
BIG-Bench Hard (maker) 94.3% ▁▁▁▁▂▄▅▆█▃▃● beats 155/155 easy · avg 78% 8.0% sharp · small run
MMLU (maker) 86.5% ▁▁▁▁▁▂▂▅██▆● beats 197/214 medium · avg 72% 7.2% small run
SimpleQA (maker) 45.9% ▄▄▄▆█▆███·●█ beats 28/32 hard · avg 17% 6.0% at its level · small run
MMLU-Pro (maker) 78.4% ▁▁·▁▁▁▃▂▄●█▁ beats 117/188 medium · avg 59% 5.9% small run
C-Eval (maker) 90.6% ▁···▁▁▂▁▂▂▃● beats 26/33 easy · avg 79% 5.1% small run
CMMLU (maker) 90.2% ▅·▅····█▃▅█● beats 14/14 easy · avg 79% 4.2% small run
ChineseSimpleQA (maker) 77.1% ▄··▄▄▄··●··█ beats 4/6 medium · avg 52% 3.5% at its level · small run
SysBench (aggregation unspecified… 69.8% █··●····█··█ beats 1/3 medium · avg 52% 3.3% at its level · small run
Multi-IF (aggregation unspecified… 76.6% ▁▁▁▁▁▁▃▄●▄▁▂ beats 23/41 medium · avg 61% 2.6% at its level · small run
MuSR (maker) 69.9% █·██······█● beats 4/4 medium · avg 57% 2.0% blunt · at its level · small run
source results share of the score
ERNIE 4.5 post-trained 300B-A47B benchm… 18 99% ▇▇▇▇▇▇▇▇▇▇▇▇
prior: pull toward the middle 1% ▇
weighted average = 60.1
Licences: ERNIE 4.5 post-trained 300B-A47B benchmark chart ([Apache-2.0](../../licenses/cards-final-NOTICE.txt))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run ERNIE 4.5 300B-A47B (ud-q4_k_xl, needs ~180.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/ERNIE-4.5-300B-A47B-PT-GGUF/tree/main))
```sh
llama cli -hf unsloth/ERNIE-4.5-300B-A47B-PT-GGUF:UD-Q4_K_XL
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf unsloth/ERNIE-4.5-300B-A47B-PT-GGUF:UD-Q4_K_XL
```
Other versions: ud-q2_k_xl (~90.5 GB) `llama cli -hf unsloth/ERNIE-4.5-300B-A47B-PT-GGUF:UD-Q2_K_XL`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run ERNIE 4.5 300B-A47B (ud-q4_k_xl, needs ~180.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/ERNIE-4.5-300B-A47B-PT-GGUF/tree/main))
```sh
lms get https://huggingface.co/unsloth/ERNIE-4.5-300B-A47B-PT-GGUF@UD-Q4_K_XL -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: ud-q2_k_xl (~90.5 GB) `lms get https://huggingface.co/unsloth/ERNIE-4.5-300B-A47B-PT-GGUF@UD-Q2_K_XL -y` · 4bit (~168.5 GB, Mac only) `lms get https://huggingface.co/mlx-community/ERNIE-4.5-300B-A47B-PT-4bit -y`
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run ERNIE 4.5 300B-A47B (4bit, needs ~168.5 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/ERNIE-4.5-300B-A47B-PT-4bit/tree/2ae73139f81d4c7ae52fc5060cc3dce46f9d0fec))
```sh
mlx_lm.chat --model mlx-community/ERNIE-4.5-300B-A47B-PT-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model mlx-community/ERNIE-4.5-300B-A47B-PT-4bit
```
Model repository and documentation: [huggingface.co/baidu](https://huggingface.co/baidu/ERNIE-4.5-300B-A47B-PT)
[interactive version](https://whatmodel.sh/model/ernie-4-5-300b-a47b) · [JSON](https://whatmodel.sh/api?format=json&all=1)