whatmodel.sh / Gemma 2B IT

Gemma 2B IT

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

Gemma 2B IT is an open-weights language model from Google, released February 2024. Its published size is 2.51B parameters and a 8k-token context window. Licence: gemma.

Catalogue identity notes

Arena entry 'gemma-2b-it' (text). LMArena's leaderboard table (arena-leaderboard space @6f2eeeb) labels 'gemma-2b-it' as 'Gemma-2B-it' (Google, Gemma license), linking https://huggingface.co/google/gemma-2b-it. Maker repo google/gemma-2b-it @9698841: licence tag gemma. 2,506,172,416 parameters from the Hugging Face safetensors metadata (dense, so active = total). Released 2024-02: Google Gemma releases page (ai.google.dev/gemma/docs/releases): 'February 21, 2024 Initial release of Gemma in 2B and 7B sizes'. The ollama gemma library was updated to v1.1; the tags used here are the separate original (non-v1.1) builds, which have different digests from the -v1.1- tags.

26.6±0.9#664 of 719 LLMsfrom 12 benchmark results across 9 sources

$ how it compares

The LLMs ranked either side of Gemma 2B IT, out of 828 with a score, then more from Google.

#modelscorewhereRAMreleased
764Flan-T5 Large Google27±2local~4.8GB2022-10
765MPT-7B-Chat MosaicML27±2comparison only2023-05
766Gemma 2B IT Google27±1local~3.1GB2024-02
767EuroLLM 1.7B Instruct UTTER (EuroLLM)27±3local~2.4GB2024-08
768Qwen2.5 0.5B Instruct Alibaba27±2local~1.8GB2024-09
more from Google
9Gemini 4 Argon (limited preview) Google82±1limited preview2026-09
23Gemini 3.8 Flash Google77±1api $1.50/M2026-09
32Gemini 3.7 Flash Google76±1api $1.50/M2026-08

see all LLMs · find the ones that fit your machine

$ what it's good at

coding▁▁▃▃▆██▇▆▄▃▁ #445 of 458 · 72% of its score · 5 tests

Beats Molmo 7B-D (0924), DeepSeek R1 Distill Qwen 1.5B, Qwen3 0.6B here, though they rank higher overall. Trails Falcon3 1B Instruct here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
HumanEval17.7% ✎▁·▁▃▃▄▆▅█▃▁▁ beats 5/99easy · avg 79%▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 35%sharp
HumanEval+15.2% ✎▁·▁▁▃▃▄▄▄█▃▁ beats 5/99easy · avg 74%▇▇▇▇▇▇▇▇▇ 23%sharp
HumanEvalSynthesize (HumanEvalPack, 6-language average)11.5% ✎▃·▂▁▂▆▅█▄▄▁▂ beats 0/35medium · avg 64%▇▇▇ 7.7%typical
HumanEvalExplain (HumanEvalPack, 6-language average)10.2% ✎▄▇█▃▄▇▄▁▇▃·▁ beats 0/35medium · avg 49%▇ 3.3%typical
HumanEvalFix (HumanEvalPack, 6-language average)16.7% ✎▁·▁▁▄██▄█▁▅▅ beats 16/35medium · avg 56%▇ 2.1%sharp · too hard for it
chat & writing▁▁▃▃▅▄▅█▆▆▇▃ #381 of 393 · 19% of its score · 4 tests

Beats Phi-2 (2.7B), OLMo 7B Instruct, Falcon 40B Instruct here, though they rank higher overall. Trails Gemma 3 270M IT, SmolLM2 360M Instruct, Granite 4.0 350M here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024)995.0 ✎▂▁▂▁▃▇█▆▇▄▂▂ beats 14/95avg 1201▇▇▇ 6.5%blunt
LMArena (text)1093.6 ✎▁▁▃▂▃▃▃▆▆▇█▃ beats 15/379avg 1268▇▇▇ 6.4%blunt
AlpacaEval 2 length-controlled5.4% ✎▁▂▃▃▆▆█▆▄▄▁▁ beats 9/122medium · avg 39%▇▇ 5.2%too hard for it
Arena-Hard (v0.1, GPT-4-1106 judge)3.0% ✎▃▁▁▅█▅▄▃▃▅▃▁ beats 0/60medium · avg 52%▇ 1.3%sharp · too hard for it
other tests3 not in a task · 8.6% of its score
testresultvs the others in its rundifficultycountswhy
WildBench v2 WB-Score (task macro)-9.6919 WB-Score ✎▃·▁▁·▃█▄▇▆▅▆ beats 0/52avg 45.199 WB-Score▇▇▇ 7.3%sharp
Open PT-LLM Leaderboard average (9 Portuguese tasks)34.3% ✎▃▁▃▁▁▁▁▁▂█▅▁ beats 40/276medium · avg 73%▇ 0.9%noisy
EuroEval HellaSwag (English), test split, 5-shot generative MC, accuracy27.6% ✎▃·▃▁▂▁▂▅█▇▄▁ beats 10/100medium · avg 68%▇ 0.4%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 Gemma 2B IT · needs ~1.5 GB · from huggingface.co/second-state
    chat in the terminal
    llama cli -hf second-state/Gemma-2b-it-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf second-state/Gemma-2b-it-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 Gemma 2B IT · needs ~1.5 GB · from huggingface.co/second-state

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/second-state/Gemma-2b-it-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 Gemma 2B IT · needs ~2.01 GB · from ollama.com
    chat in the terminal
    ollama run gemma:2b-instruct-q4_K_M

Official weights and documentation: huggingface.co/google ↗

$ how the score was worked out · EvalPlus / EvalPerf published… 58%, Granite Code Models: A Family… 13%, WildBench v2 leaderboard data… 7% and 6 more

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
EvalPlus / EvalPerf published evaluation runs2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 58%Apache-2.0
Granite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runs3▇▇▇▇ 13%CC BY 4.0
WildBench v2 leaderboard data (Ai2)1▇▇ 7%Apache-2.0
Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)1▇▇ 7%CC-BY-4.0
LMArena leaderboard dataset1▇▇ 6%CC BY 4.0
AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)1▇▇ 5%Apache-2.0
Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)1▇ 1%Apache-2.0
Open Portuguese LLM Leaderboard (archived)1▇ 1%Apache-2.0
EuroEval leaderboards (results collection), English generative leaderboard1▇ 0%MIT
prior · pull toward the middle▇ 0%
weighted average= 26.6

$ 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