[whatmodel.sh](../../README.md) / Code Tulu 2 7B # Code Tulu 2 7B Allen Institute for AI · released 2023-11 · 7B params · AI2 ImpACT Low-risk; Llama 2 base terms **score 39.2** ±1.7 · #576 of 732 · from 5 benchmark results across 1 source · † no independently run results verified Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at _{● marks Code Tulu 2 7B among every model on that task or test, worst to best.}

task    vs the others          rank  counts  tests
coding  ▁▁▁▃●▇█▇▅▄▃▁   #388 of 469*     19%      1
math    ▁▁▃▅●▆▄█▄▂▁▁   #321 of 431*     21%      1

coding · #388 of 469 · 19% of its score

Beats [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md), [DeepSeek R1 Distill Qwen 7B](deepseek-r1-distill-qwen-7b.md), [Tulu 2 70B](tulu-2-70b.md) here, though they rank higher overall. Trails [Qwen2.5 1.5B Instruct](qwen2-5-1-5b.md), [WizardLM 13B V1.2](wizardlm-13b-v1-2.md), [Falcon3 3B Instruct](falcon3-3b.md) here, though they rank lower overall.

test                               result  vs the others in its run  difficulty      counts  why
Tulu 2: CodexEval pass@10 (maker)   68.9%  ▃▃▃·▆▃▃●▃▂▂▅ beats 15/25  easy · avg 83%     19%  sharp · at its level

math · #321 of 431 · 21% of its score

Beats [GPT-4 (0613)](gpt-4-0613.md), [Athene-70B](athene-70b.md), [Gemma 2 27B IT](gemma-2-27b.md) here, though they rank higher overall. Trails [StarCoder2 7B](starcoder2-7b.md), [DeepSeek R1 Distill Qwen 1.5B](deepseek-r1-distill-qwen-1-5b.md), [Tulu 2 DPO 7B](tulu-2-dpo-7b.md) here, though they rank lower overall.

test           result  vs the others in its run  difficulty      counts  why
GSM8K (maker)   33.0%  ▆▂▅●▅▅▃▆···▂ beats 9/25   easy · avg 72%     21%  sharp · at its level

other tests (3)

test                                result  vs the others in its run  difficulty        counts  why
BIG-Bench Hard (maker)               49.1%  ▁·····▂●▂▅·▁ beats 11/25  easy · avg 74%       38%  sharp · at its level
MMLU (maker)                         43.7%  ▁▁▃●▄█▄▁█··▁ beats 5/25   medium · avg 73%     17%  at its level
Tulu 2: TyDiQA gold passage, 1-sh…   52.6%  ▂··▁▃▂▂▂●▅▁▂ beats 16/25  medium · avg 59%    4.3%  blunt · at its level

## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's 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 is added up by who ran the tests. [Method](../method.md)

source                               results  share of the score
Tulu 2 (submitted / maker-reported)        5   99% ▇▇▇▇▇▇▇▇▇▇▇▇
prior: pull toward the middle                   1% ▇
weighted average                              = 39.2

_{Licences: Tulu 2 ([CC BY 4.0](../../licenses/CC-BY-4.0.txt))} ## How to run it ### Other ways to run it 1. Run it as published (BF16, needs ~13.48 GB, from [huggingface.co/allenai](https://huggingface.co/allenai/codetulu-2-7b/tree/99a3a11ee3320d52eae6cf5920c80e1d51d19a6e)) ```sh uv run --with torch --with 'transformers<5' --with accelerate python -c "from transformers import pipeline; print(pipeline('text-generation', model='allenai/codetulu-2-7b', torch_dtype='auto', device_map='auto')('Once upon a time', max_new_tokens=40)[0]['generated_text'])" ``` Model repository and documentation: [huggingface.co/allenai](https://huggingface.co/allenai/codetulu-2-7b) [interactive version](https://whatmodel.sh/model/code-tulu-2-7b) · [JSON](https://whatmodel.sh/api?format=json&all=1)