whatmodel.sh / StableLM-3B-4E1T

StableLM-3B-4E1T

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

StableLM-3B-4E1T is an open-weights language model from Stability AI, released September 2023. Its published size is 2.8B parameters and a 4k-token context window. Licence: cc-by-sa-4.0.

Catalogue identity notes

3B base model trained on 1T tokens for 4 epochs; not StableLM Zephyr or StableLM 2.

28.9±0.9#647 of 719 LLMsfrom 11 benchmark results across 2 sources

$ how it compares

The LLMs ranked either side of StableLM-3B-4E1T, out of 828 with a score, then more from Stability AI.

#modelscorewhereRAMreleased
744MPT-7B (base) MosaicML30±1comparison only2023-05
745OLMo 7B Instruct Ai229±2local~16GB2024-02
746StableLM-3B-4E1T Stability AI29±1local~3.3GB2023-09
747Gemma 2B (base) Google29±1local~3.2GB2024-02
748OpenLLaMA 7B v2 OpenLM Research29±1local~9.1GB2023-07
more from Stability AI
718Stable Code 3B Stability AI33±1local~3.3GB2024-01
750StableLM-Base-Alpha 7B v2 Stability AI29±1local~30GB2023-08
785StableLM-Base-Alpha 3B v2 Stability AI24±1local~13GB2023-08

see all LLMs · find the ones that fit your machine

$ what it's good at

other tests11 not in a task · 100% of its score
testresultvs the others in its rundifficultycountswhy
ARC Challenge lab40.0% ✎▁·▁··▁▁▇▂█▆▃ beats 13/26easy · avg 85%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 46%sharp
WinoGrande lab66.5% ✎▁····▁··▁▂█▄ beats 16/26easy · avg 89%▇▇▇▇▇ 12%at its level
BoolQ lab75.6% ✎▂▃▂▅▆▂▅▆█▂▂▂ beats 23/26easy · avg 91%▇▇▇▇ 10%at its level
ARC Easy lab67.7% ✎▁▁·▁▁▄▅▄█▅▁▃ beats 14/26easy · avg 88%▇▇▇ 7.0%at its level
HellaSwag lab73.9% ✎▃···▁▇▅█▇▃▁▁ beats 16/26very easy · avg 92%▇▇ 5.7%at its level
SciQ lab94.8% ✎▁·▁·▁·▃▅▃█▄▁ beats 25/26very easy · avg 98%▇▇ 5.4%typical
PIQA lab79.2% ✎▁·▁··▁▁▃█▅█▂ beats 24/26easy · avg 88%▇▇ 4.4%blunt · at its level
OpenBookQA lab31.4% ✎█·▁▁▃·▁▃▅▇▃▄ beats 20/26medium · avg 53%▇ 3.5%too hard for it
3 more tests
testresultvs the others in its rundifficultycountswhy
LAMBADA (OpenAI version), last-word accuracy lab70.6% ✎▁··▂▁▁▁▂▆█▅▃ beats 16/26very easy · avg 89%▇ 2.7%blunt · at its level
Social IQA lab41.9% ✎▁···▁▃█▁▁▁·▁ beats 9/26medium · avg 59%▇ 2.5%blunt
Open PT-LLM Leaderboard average (9 Portuguese tasks)39.8% ✎▃▁▁▁▁▃▁▁▂█▅▁ beats 49/276medium · avg 73%▇ 1.2%noisy

$ where to run it

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run StableLM-3B-4E1T · needs ~1.7 GB · from huggingface.co/afrideva
    chat in the terminal
    llama cli -hf afrideva/stablelm-3b-4e1t-GGUF:q4_k_m
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf afrideva/stablelm-3b-4e1t-GGUF:q4_k_m

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run StableLM-3B-4E1T · needs ~1.7 GB · from huggingface.co/afrideva

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/afrideva/stablelm-3b-4e1t-GGUF@q4_k_m -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run StableLM-3B-4E1T · needs ~1.7 GB · from huggingface.co/afrideva
    chat in the terminal
    ollama run hf.co/afrideva/stablelm-3b-4e1t-GGUF:q4_k_m

Other ways to run it

  1. Run it as published · needs ~5.59 GB · from huggingface.co/stabilityai (official)
    uv run --with torch --with 'transformers<5' python -c "from transformers import pipeline; print(pipeline('text-generation', model='stabilityai/stablelm-3b-4e1t')('Once upon a time', max_new_tokens=40)[0]['generated_text'])"

Official weights and documentation: huggingface.co/stabilityai ↗

$ how the score was worked out · Stability AI — original Stabl… 98%, Open Portuguese LLM Leaderboa… 1%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Stability AI — original StableLM zero-shot evaluation outputs · lab-reported10▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 98%Apache-2.0
Open Portuguese LLM Leaderboard (archived)1▇ 1%Apache-2.0
prior · pull toward the middle▇ 0%
weighted average= 28.9

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json