whatmodel.sh / Code Tulu 2 13B

Code Tulu 2 13B

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

Code Tulu 2 13B is an open-weights language model from Allen Institute for AI, released November 2023. Its published size is 13B parameters. Licence: AI2 ImpACT Low-risk; Llama 2 base terms.

Catalogue identity notes

Exact full-precision model evaluated in the Tulu 2 paper (arXiv:2311.10702v2). QLoRA ablations are separate and not imported.

Model information sources

Release date ↗ · Weights licence ↗

39.6±1.7#536 of 719 LLMsfrom 5 benchmark results across 1 source

$ how it compares

The LLMs ranked either side of Code Tulu 2 13B, out of 828 with a score, then more from Allen Institute for AI.

#modelscorewhereRAMreleased
626Starling-LM 7B beta Nexusflow40±1local~6GB2024-03
627LFM2 1.2B Liquid AI40±2local~2.2GB2025-07
628Code Tulu 2 13B Allen Institute for AI40±2local~29GB2023-11
629LLaVA-v1.6-34B LLaVA40±2local~24GB2024-01
630Bielik 1.5B v3.0 Instruct SpeakLeash40±3local~2.4GB2025-04
more from Allen Institute for AI
493OLMo 2 13B Instruct (1124) Allen Institute for AI46±1local~9.6GB2024-12
516Tulu 2 70B Allen Institute for AI46±2local~28GB2023-11
557Code Tulu 2 34B Allen Institute for AI44±2local~72GB2023-11

see all LLMs · find the ones that fit your machine

$ what it's good at

coding▁▁▁▃▆██▇▆▄▃▁ #346 of 458 * · 21% of its score · 1 test

Beats LFM2.5 2.6B, LFM2.5 8B-A1B, Qwen3-VL 2B Thinking here, though they rank higher overall. Trails Phi-3-mini-128k-instruct (Apr 2024) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Tulu 2: CodexEval pass@10 lab76.2% ✎▃▃▃·▆▃▃█▃▂▂▅ beats 19/25very easy · avg 86%▇▇▇▇▇▇▇▇ 21%sharp · at its level
math▁▂▃▅▆▅▄█▄▁▁▁ #332 of 398 * · 14% of its score · 1 test

Beats GPT-3.5 Turbo (0125), Qwen2-VL 7B Instruct, DBRX Instruct here, though they rank higher overall. Trails Qwen3 0.6B, Gemma 7B (base), DeepSeek R1 Distill Qwen 1.5B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
GSM8K lab41.0% ✎▆▂▅█▅▅▃▆···▂ beats 14/25easy · avg 73%▇▇▇▇▇▇ 14%sharp · at its level
other tests3 not in a task · 63% of its score
testresultvs the others in its rundifficultycountswhy
BIG-Bench Hard lab52.8% ✎▁·····▂█▃▅·▁ beats 15/25easy · avg 77%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 43%sharp · at its level
MMLU lab45.9% ✎▁▁▃▃▄█▄▁█··▁ beats 6/25easy · avg 74%▇▇▇▇▇▇ 15%at its level
Tulu 2: TyDiQA gold passage, 1-shot F1 lab55.7% ✎▂··▁▃▂▂▂█▅▁▂ beats 20/25medium · avg 61%▇▇ 4.9%blunt · at its level

$ where to run it

Other ways to run it

  1. Run it as published · needs ~26.03 GB · from huggingface.co/allenai
    uv run --with torch --with 'transformers<5' --with accelerate python -c "from transformers import pipeline; print(pipeline('text-generation', model='allenai/codetulu-2-13b', torch_dtype='auto', device_map='auto')('Once upon a time', max_new_tokens=40)[0]['generated_text'])"

Weights and documentation: huggingface.co/allenai ↗

$ how the score was worked out · Tulu 2 — base and instruction… 99%

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
Tulu 2 — base and instruction-tuned evaluations · lab-reported5▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 99%CC BY 4.0
prior · pull toward the middle▇ 1%
weighted average= 39.6

$ 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