[whatmodel.sh](../../README.md) / Pythia 410M
# Pythia 410M
EleutherAI · released 2023-02 · 0.405B params · apache-2.0
**score 7.6** ±1.8 · #720 of 732 · from 8 benchmark results across 3 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Pythia 410M among every model on that task or test, worst to best.
other tests (8)
test result vs the others in its run difficulty counts why
ARC Easy (maker) 52.1% ▃·▅▃▃●·██▃██ beats 6/19 easy · avg 87% 21% blunt · at its level
SciQ (maker) 81.1% ▂··▂▃▂●▆▃▃█▂ beats 5/19 very easy · avg 98% 19% blunt
ImageNetVC visual commonsense, te… 37.1% ▂·▄·●█▄▄█▄▂▆ beats 3/21 easy · avg 74% 16% blunt
PIQA (maker) 66.8% ▂·▄▄·●▂▆▂▄█▄ beats 5/19 easy · avg 87% 13% blunt · at its level
LAMBADA (OpenAI version), last-wo… 51.6% ▃··▅▃▃▅●▅███ beats 7/19 very easy · avg 87% 12% blunt · at its level
WinoGrande (maker) 53.7% ▄····▄●▂▄▆▆█ beats 6/19 easy · avg 87% 11% typical
ARC Challenge (maker) 21.3% ●··▁·▁▁·▁▂▁▂ beats 7/19 medium · avg 71% 6.3% sharp · too hard for it
Open PT-LLM Leaderboard average (… 26.4% ●▁▁▁▁▁▁▁▂█▅▁ beats 7/276 medium · avg 72% 1.1% too hard for it · noisy
source results share of the score
Pythia: A Suite for Analyzing Large Lan… 6 82% ▇▇▇▇▇▇▇▇▇▇
ImageNetVC: Zero- and Few-Shot Visual C… 1 16% ▇▇
Open Portuguese LLM Leaderboard 1 1% ▇
prior: pull toward the middle 1% ▇
weighted average = 7.6
Licences: Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling ([CC BY-SA 4.0](../../licenses/paper-2304.01373v2-CC-BY-SA-4.0.txt)) · ImageNetVC: Zero- and Few-Shot Visual Commonsense Evaluation on 1000 ImageNet Categories ([CC BY 4.0](../../licenses/acl-2023-findings-emnlp-133-CC-BY-4-0.txt)) · 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 Pythia 410M (Q4_K_M, needs ~0.3 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/pythia-410m-GGUF/tree/29569dd6a296a653472b3086dbc8ded8479e9707))
```sh
llama cli -hf mradermacher/pythia-410m-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf mradermacher/pythia-410m-GGUF:Q4_K_M
```
Other versions: Q8_0 (~0.4 GB) `llama cli -hf mradermacher/pythia-410m-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 Pythia 410M (Q4_K_M, needs ~0.3 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/pythia-410m-GGUF/tree/29569dd6a296a653472b3086dbc8ded8479e9707))
```sh
lms get https://huggingface.co/mradermacher/pythia-410m-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/pythia-410m-GGUF@Q8_0 -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 Pythia 410M (Q4_K_M, needs ~0.3 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/pythia-410m-GGUF/tree/29569dd6a296a653472b3086dbc8ded8479e9707))
```sh
ollama run hf.co/mradermacher/pythia-410m-GGUF:Q4_K_M
```
Other versions: Q8_0 (~0.4 GB) `ollama run hf.co/mradermacher/pythia-410m-GGUF:Q8_0`
### Other ways to run it
1. Run it as published (fp16, needs ~0.91 GB, from [huggingface.co/EleutherAI](https://huggingface.co/EleutherAI/pythia-410m/tree/9879c9b5f8bea9051dcb0e68dff21493d67e9d4f))
```sh
uv run --with torch --with 'transformers<5' python -c "from transformers import pipeline; print(pipeline('text-generation', model='EleutherAI/pythia-410m')('Once upon a time', max_new_tokens=40)[0]['generated_text'])"
```
Model repository and documentation: [huggingface.co/EleutherAI](https://huggingface.co/EleutherAI/pythia-410m)
[interactive version](https://whatmodel.sh/model/pythia-410m) · [JSON](https://whatmodel.sh/api?format=json&all=1)