[whatmodel.sh](../../README.md) / Gemma 2B (base)
# Gemma 2B (base)
Google · released 2024-02 · 2.51B params · gemma
**score 30.9** ±0.6 · #656 of 732 · from 37 benchmark results across 9 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Gemma 2B (base) among every model on that task or test, worst to best.
task vs the others rank counts tests
coding ▁▁▁●▅▇█▇▅▄▃▁ #447 of 469 67% 7
math ▁▁▃▅●▆▄█▄▂▁▁ #334 of 431* 3.9% 2
coding · #447 of 469 · 67% of its score
Beats [Qwen3.5 2B](qwen3-5-2b.md), [Molmo 7B-D (0924)](molmo-7b-d.md), [Tulu 2 7B](tulu-2-7b.md) here, though they rank higher overall. Trails [Falcon3 1B Instruct](falcon3-1b.md), [Granite 3.0 1B-A400M Instruct](granite-3-0-1b-a400m.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
HumanEval 25.0% ▁·▁▁●▄▆▅█▃▁▁ beats 9/99 easy · avg 76% 37% sharp · several runs
HumanEval+ 20.7% ▁·▁▁▁●▄▄▄█▃▁ beats 8/99 easy · avg 70% 14% sharp
HumanEvalSynthesize (HumanEvalPac… 17.2% ▁·●▁▂▆▅█▄▄▁▂ beats 1/35 medium · avg 62% 4.6% typical
MBPP+ 34.1% ▁···▁●▂▅▇█▆▂ beats 2/75 easy · avg 66% 4.1% at its level
MBPP (maker) 29.2% ▁▁●▄▅█▆▅▄▁▂▁ beats 7/86 easy · avg 71% 3.3% at its level · several runs
HumanEvalExplain (HumanEvalPack, … 10.9% ●▇█▃▄▇▄▁▇▃·▁ beats 1/35 medium · avg 46% 2.1% typical
HumanEvalFix (HumanEvalPack, 6-la… 1.9% ●·▁▁▄██▄█▁▅▅ beats 0/35 medium · avg 51% 1.5% sharp
math · #334 of 431 · 3.9% of its score
Beats [GPT-4 (0613)](gpt-4-0613.md), [Athene-70B](athene-70b.md), [Llama 3 70B Instruct](llama-3-70b.md) here, though they rank higher overall.
test result vs the others in its run difficulty counts why
GSM8K (maker) 17.7% ▂▂▂▂▃●▅▇██▆▄ beats 48/206 easy · avg 78% 2.6% sharp · small run
MATH (maker) 11.8% ▃▅▂●▅▃▅▅█▃▃▂ beats 9/30 medium · avg 49% 1.3% small run
other tests (22)
test result vs the others in its run difficulty counts why
EvoEval (overall) 12.8% ●▂▃▄▄█▁▄▅▃▁▃ beats 0/42 medium · avg 52% 5.6% typical
Natural Questions (closed book) 14.8% ▁·▁·▁▂▁▂·●▂█ beats 13/28 medium · avg 48% 3.6% several runs
CommonsenseQA (maker) 65.3% █·····▂▂▃▂●▄ beats 27/38 very easy · avg 94% 2.9% sharp · at its level · small run
BIG-Bench Hard (maker) 35.2% ▁▁▁▁▂●▅▆█▃▃▁ beats 24/155 easy · avg 79% 2.0% sharp · small run
MedQA (USMLE, 4 options) 29.2% ▁▁··▄●▁··▄▃█ beats 5/17 medium · avg 71% 1.9% sharp
ARC Challenge (maker) 42.1% ▄▁▁▂▁▄●▇▃▂▃▂ beats 59/140 easy · avg 78% 1.8% sharp · small run
TriviaQA 47.8% ▁··▁█▃▃▁▄●▄█ beats 19/28 easy · avg 81% 1.4% at its level · small run · several runs
MMLU (maker) 42.3% ▁▁▁▁▁▂●▅██▆▄ beats 33/214 easy · avg 81% 1.4% at its level · small run
OpenBookQA (maker) 47.8% ▅▁▁▁▂▂▃●▃▁▃▁ beats 62/107 medium · avg 68% 1.1% small run
WinoGrande (maker) 65.4% ▁▁▁▁▂▁▁▅●█▃▁ beats 62/131 easy · avg 84% 1.0% at its level · small run
MedMCQA 30.2% ▂▂··●▄▂▂·▄██ beats 2/17 medium · avg 63% 1.0% typical
AGIEval English (Gemma report) (m… 24.2% ●··········█ beats 0/2 medium · avg 62% 0.9% at its level · small run
HellaSwag (maker) 71.4% ▁▁▁▂▂▂▄▆●▇▄▂ beats 72/135 easy · avg 89% 0.9% at its level · small run
BoolQ (maker) 69.4% ▁····▁▂█●▅▆▄ beats 50/103 easy · avg 89% 0.8% at its level · small run
Open PT-LLM Leaderboard average (… 44.1% ▃▁▁▁▁▁●▁▂█▅▁ beats 54/276 medium · avg 72% 0.6% noisy
ARC Easy (maker) 73.2% ▁▃▂▄▆█●▃▁··▁ beats 74/100 easy · avg 87% 0.5% blunt · at its level · small run
Social IQA (maker) 49.7% ▁·▁██▄●▁▂▃▇▁ beats 34/59 medium · avg 63% 0.4% blunt · small run
EuroEval HellaSwag (English), tes… 26.9% ▃·●▁▂▁▂▅█▇▄▁ beats 9/100 medium · avg 67% 0.3% sharp · noisy
PIQA (maker) 77.3% ▁▁▁▂▂▃▅●▅▃▁▁ beats 66/123 easy · avg 86% 0.3% blunt · at its level · small run
PubMedQA 66.4% ▂·▂▆·▄●▄▄▂█▂ beats 7/17 medium · avg 75% 0.2% blunt · at its level
HaluEval QA hallucination detecti… 48.6% ●▁▁·▁▁·▁▁▁▁▁ beats 7/27 hard · avg 66% 0.2% too hard for it · noisy
TruthfulQA MC2 33.1% ●▃█▆▁▂▅···▁▁ beats 0/31 medium · avg 57% 0.1% blunt · at its level
source results share of the score
EvalPlus / EvalPerf published evaluatio… 3 39% ▇▇▇▇▇
Gemma report (submitted / maker-reporte… 18 23% ▇▇▇
Top Leaderboard Ranking = Top Coding Pr… 2 17% ▇▇
Granite Code Models: A Family of Open F… 3 8% ▇
CodeGemma: Open Code Models Based on Ge… 2 5% ▇
Hallucinations Leaderboard results 4 3% ▇
Open Medical-LLM Leaderboard results 3 3% ▇
Open Portuguese LLM Leaderboard 1 1% ▇
EuroEval leaderboards 1 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 30.9
Licences: EvalPlus / EvalPerf published evaluation runs ([Apache-2.0](../../licenses/evalplus-published-runs-Apache-2.0.txt)) · Gemma report ([CC BY 4.0](../../licenses/CC-BY-4.0.txt)) · Top Leaderboard Ranking = Top Coding Proficiency ([CC BY 4.0](../../licenses/paper-2403.19114v1-CC-BY-4.0.txt)) · Granite Code Models: A Family of Open Foundation Models for Code Intelligence ([CC BY 4.0](../../licenses/paper-2405.04324v1-CC-BY-4.0.txt)) · CodeGemma: Open Code Models Based on Gemma ([CC BY 4.0](../../licenses/paper-2406.11409v2-CC-BY-4.0.txt)) · Hallucinations Leaderboard results ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)) · Open Medical-LLM Leaderboard results ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE))
## 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 Gemma 2B (base) (Q4_K_M, needs ~1.6 GB, from [huggingface.co/MaziyarPanahi](https://huggingface.co/MaziyarPanahi/gemma-2b-GGUF/tree/e08e4cde129cab50844282f4078d7b9d5a76a36d))
```sh
llama cli -hf MaziyarPanahi/gemma-2b-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf MaziyarPanahi/gemma-2b-GGUF:Q4_K_M
```
Other versions: Q8_0 (~2.7 GB) `llama cli -hf MaziyarPanahi/gemma-2b-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 Gemma 2B (base) (Q4_K_M, needs ~1.6 GB, from [huggingface.co/MaziyarPanahi](https://huggingface.co/MaziyarPanahi/gemma-2b-GGUF/tree/e08e4cde129cab50844282f4078d7b9d5a76a36d))
```sh
lms get https://huggingface.co/MaziyarPanahi/gemma-2b-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 (~2.7 GB) `lms get https://huggingface.co/MaziyarPanahi/gemma-2b-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 Gemma 2B (base) (Q4_K_M, needs ~1.8 GB, from [ollama.com](https://ollama.com/library/gemma/tags))
```sh
ollama run gemma:2b-text-q4_K_M
```
Other versions: Q8_0 (~2.7 GB) `ollama run gemma:2b-text-q8_0`
Model repository and documentation: [huggingface.co/google](https://huggingface.co/google/gemma-2b)
[interactive version](https://whatmodel.sh/model/gemma-2b-base) · [JSON](https://whatmodel.sh/api?format=json&all=1)