[whatmodel.sh](../../README.md) / GPT-2 Medium (355M)
# GPT-2 Medium (355M)
OpenAI · released 2019-05 · 0.355B params · mit
**score -5.4** ±1.3 · #730 of 732 · from 36 benchmark results across 9 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks GPT-2 Medium (355M) among every model on that task or test, worst to best.
task vs the others rank counts tests
math ●▁▃▅█▆▄█▄▂▁▁ #430 of 431* 0.3% 1
math · #430 of 431 · 0.3% of its score
Beats [GPT-2 Large (774M)](gpt-2-large.md) here, though they rank higher overall. Trails [GPT-2 Small (124M)](gpt-2-small.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
GSM8K 0.0% ●·▂█·······▂ beats 1/9 medium · avg 62% 0.3% sharp · too hard for it · noisy
other tests (20)
test result vs the others in its run difficulty counts why
OpenBookQA 18.6% ●▁▁▁▂▂▃█▃▁▃▁ beats 12/107 very hard · avg 26% 20% too hard for it · small run · several r…
ARC Challenge 22.8% ●▁▁▂▁▄█▇▃▂▃▂ beats 4/140 very hard · avg 27% 18% sharp · too hard for it · several runs
HellaSwag 32.2% ▁▁▁●▂▂▄▆█▇▄▂ beats 9/135 easy · avg 87% 13% several runs
COPA 67.0% ▁·▁▁▁●▅█▆▂▁▁ beats 15/67 very easy · avg 97% 13% at its level · several runs
Story Cloze Test 63.0% ▃▃●▂▅█▅▂▁··▂ beats 12/50 easy · avg 91% 10% blunt · several runs
ARC Easy 49.4% ▁▃●▄▆█▅▃▁··▁ beats 12/100 very easy · avg 90% 9.3% blunt · at its level · several runs
PROST physical reasoning (zero-sh… 26.8% █···██·●···█ beats 3/4 easy · avg 86% 4.8% too hard for it
COPEN concept identification in c… 32.0% ▃···▃●·▃·▃·█ beats 2/7 easy · avg 81% 3.5% too hard for it
Social IQA 41.4% ▁·▁●█▄▅▁▂▃▇▁ beats 12/59 medium · avg 71% 2.2% blunt
PIQA 67.6% ▁▁▁●▂▃▅█▅▃▁▁ beats 15/123 easy · avg 89% 1.9% blunt · at its level
BoolQ 56.6% ▁····▁●█▆▅▆▄ beats 4/103 easy · avg 91% 1.6% noisy
COPEN conceptual similarity judgm… 10.5% ▄···█···●█·▄ beats 4/7 hard · avg 22% 0.5% blunt · too hard for it
SVAMP 0.0% ●·▆▂▄▂▂····▂ beats 1/12 easy · avg 70% 0.4% sharp · too hard for it
StrategyQA 0.1% ●········▁▃▁ beats 1/9 hard · avg 62% 0.3% too hard for it · noisy
CommonsenseQA 0.0% ●····▁·····▁ beats 1/9 easy · avg 75% 0.3% sharp · too hard for it · several runs
Open PT-LLM Leaderboard average (… 29.1% ●▁▁▁▁▁▁▁▂█▅▁ beats 24/276 medium · avg 72% 0.2% too hard for it · noisy
EuroEval Life in the UK (English … 25.9% ▁●▁▁▁▁▁▂▄█▄▃ beats 2/146 easy · avg 82% 0.2% too hard for it · noisy
EuroEval HellaSwag (English), tes… 24.9% ●·▁▁▂▁▂▅█▇▄▁ beats 3/100 medium · avg 67% 0.0% sharp · too hard for it · noisy
COPEN conceptual property judgmen… 47.2% ●·······▁··▁ beats 5/7 very hard · avg 52% 0.0% blunt · too hard for it · noisy
AQuA-RAT 3.9% ●·····▁·▁··▁ beats 1/9 medium · avg 53% 0.0% sharp · too hard for it · noisy
source results share of the score
Coherence boosting: When your pretraine… 9 35% ▇▇▇▇
Surface Form Competition 7 25% ▇▇▇
CASE: Commonsense-Augmented Score with … 6 24% ▇▇▇
RankGen: Improving Text Generation with… 2 5% ▇
PROST 1 5% ▇
COPEN 3 4% ▇
Large Language Models Are Reasoning Tea… 5 1% ▇
EuroEval leaderboards 2 0% ▇
Open Portuguese LLM Leaderboard 1 0% ▇
prior: pull toward the middle 1% ▇
weighted average = -5.4
Licences: Coherence boosting: When your pretrained language model is not paying enough attention ([CC BY 4.0](../../licenses/acl-2022-acl-long-565-CC-BY-4-0.txt)) · Surface Form Competition ([CC BY 4.0](../../licenses/acl-2021-emnlp-main-564-CC-BY-4.0.txt)) · CASE: Commonsense-Augmented Score with an Expanded Answer Space ([CC BY 4.0](../../licenses/acl-2023-findings-emnlp-180-CC-BY-4-0.txt)) · RankGen: Improving Text Generation with Large Ranking Models ([CC BY 4.0](../../licenses/acl-2022-emnlp-main-15-CC-BY-4-0.txt)) · PROST ([CC BY 4.0](../../licenses/_cc-by-4.0-legalcode.txt)) · COPEN ([CC BY 4.0](../../licenses/copen-emnlp2022-zero-shot-CC-BY-4.0.txt)) · Large Language Models Are Reasoning Teachers ([CC BY 4.0](../../licenses/acl-2023-acl-long-830-CC-BY-4-0.txt)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.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 GPT-2 Medium (355M) (Q4_K_M, needs ~0.3 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/gpt2-medium-GGUF/tree/3b9897d67a84e967fbcc8d7de3db4c797e386740))
```sh
llama cli -hf mradermacher/gpt2-medium-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf mradermacher/gpt2-medium-GGUF:Q4_K_M
```
Other versions: Q8_0 (~0.4 GB) `llama cli -hf mradermacher/gpt2-medium-GGUF:Q8_0`
### 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 GPT-2 Medium (355M) (Q4_K_M, needs ~0.3 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/gpt2-medium-GGUF/tree/3b9897d67a84e967fbcc8d7de3db4c797e386740))
```sh
lms get https://huggingface.co/mradermacher/gpt2-medium-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q8_0 (~0.4 GB) `lms get https://huggingface.co/mradermacher/gpt2-medium-GGUF@Q8_0 -y` · 8bit (~0.4 GB, Mac only) `lms get https://huggingface.co/viktor2698/gpt2-medium-mlx-8Bit -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run GPT-2 Medium (355M) (Q4_K_M, needs ~0.3 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/gpt2-medium-GGUF/tree/3b9897d67a84e967fbcc8d7de3db4c797e386740))
```sh
ollama run hf.co/mradermacher/gpt2-medium-GGUF:Q4_K_M
```
Other versions: Q8_0 (~0.4 GB) `ollama run hf.co/mradermacher/gpt2-medium-GGUF:Q8_0`
### 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 GPT-2 Medium (355M) (8bit, needs ~0.4 GB, Mac only, from [huggingface.co/viktor2698](https://huggingface.co/viktor2698/gpt2-medium-mlx-8Bit/tree/8bb3814b1fcfa9d04b3d051102b19d44b3e5064f))
```sh
mlx_lm.chat --model viktor2698/gpt2-medium-mlx-8Bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model viktor2698/gpt2-medium-mlx-8Bit
```
### Other ways to run it
1. Run it as published (fp32, needs ~1.52 GB, from [huggingface.co/openai-community](https://huggingface.co/openai-community/gpt2-medium/tree/6dcaa7a952f72f9298047fd5137cd6e4f05f41da))
```sh
uv run --with torch --with 'transformers<5' python -c "from transformers import pipeline; print(pipeline('text-generation', model='openai-community/gpt2-medium')('Once upon a time', max_new_tokens=40)[0]['generated_text'])"
```
Model repository and documentation: [huggingface.co/openai-community](https://huggingface.co/openai-community/gpt2-medium)
[interactive version](https://whatmodel.sh/model/gpt-2-medium) · [JSON](https://whatmodel.sh/api?format=json&all=1)