whatmodel.sh / OPT-6.7B

OPT-6.7B

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

OPT-6.7B is an open-weights language model from Meta, released May 2022. Its published size is 6.7B parameters and a 2k-token context window. Licence: opt-175b-license-noncommercial.

Catalogue identity notes

Meta's Open Pre-trained Transformer base model as released on Hugging Face (weights under Meta's non-commercial OPT licence).

24.3±0.8#677 of 719 LLMsfrom 17 benchmark results across 3 sources

$ how it compares

The LLMs ranked either side of OPT-6.7B, out of 828 with a score, then more from Meta.

#modelscorewhereRAMreleased
780GPT-J 6B EleutherAI25±1local~14GB2021-06
781OpenLLaMA 3B v2 OpenLM Research25±1local~4.2GB2023-07
782OPT-6.7B Meta24±1local~16GB2022-05
783RWKV-4-Raven 14B RWKV (BlinkDL)24±2comparison only2023-04
784Falcon3 1B Instruct TII24±4local~2.4GB2024-12
more from Meta
24Muse Spark 1.3 Meta77±1api $2.00/M2026-09
40Muse Spark 1.2 Meta75±1api $2.00/M2026-08
43Muse Spark 1.1 Meta75±1api $2.00/M2026-07

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 +134.8% ✎▁·▁··▁▁▇▂█▆▃ beats 6/26easy · avg 85%▇▇▇▇▇▇▇▇▇▇▇▇▇ 33%sharp · several runs
WinoGrande +165.4% ✎▁····▁··▁▂█▄ beats 12/26easy · avg 89%▇▇▇▇▇▇ 15%at its level · several runs
ARC Easy +160.1% ✎▁▁·▁▁▄▅▄█▅▁▃ beats 6/26very easy · avg 90%▇▇▇▇▇ 13%at its level · several runs
SciQ +190.1% ✎▁·▁·▁·▃▅▃█▄▁ beats 6/26very easy · avg 98%▇▇▇▇ 9.3%several runs
PIQA +176.3% ✎▁·▁··▁▁▃█▅█▂ beats 13/26easy · avg 88%▇▇▇ 7.5%blunt · at its level · several runs
BoolQ66.0% ✎▂▃▂▅▆▂▅▆█▂▂▂ beats 10/26easy · avg 91%▇▇▇ 7.2%at its level
LAMBADA (OpenAI version), last-word accuracy +167.6% ✎▁··▂▁▁▁▂▆█▅▃ beats 10/26very easy · avg 89%▇▇ 5.4%blunt · at its level · several runs
HellaSwag67.2% ✎▃···▁▇▅█▇▃▁▁ beats 9/26very easy · avg 92%▇▇ 4.9%at its level
3 more tests
testresultvs the others in its rundifficultycountswhy
OpenBookQA27.6% ✎█·▁▁▃·▃▃▅▇▃▄ beats 10/26medium · avg 53%▇ 1.9%too hard for it
Social IQA42.6% ✎▁···▁▃█▁▁▁·▁ beats 18/26medium · avg 59%▇ 1.7%blunt
Open PT-LLM Leaderboard average (9 Portuguese tasks)28.6% ✎▃▁▁▁▁▁▁▁▂█▅▁ beats 12/276medium · avg 73%▇ 0.7%noisy

$ where to run it

Other ways to run it

  1. Run it as published · needs ~13.32 GB · from huggingface.co/facebook
    uv run --with torch --with 'transformers<5' --with accelerate python -c "from transformers import pipeline; print(pipeline('text-generation', model='facebook/opt-6.7b', torch_dtype='auto', device_map='auto')('Once upon a time', max_new_tokens=40)[0]['generated_text'])"

Weights and documentation: huggingface.co/facebook ↗

$ how the score was worked out · Stability AI — original Stabl… 67%, Pythia: A Suite for Analyzing… 32%, 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 outputs10▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 67%Apache-2.0
Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runs6▇▇▇▇▇▇▇▇▇▇ 32%CC BY-SA 4.0
Open Portuguese LLM Leaderboard (archived)1▇ 1%Apache-2.0
prior · pull toward the middle▇ 0%
weighted average= 24.3

$ 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