whatmodel.sh / Gemma 3 1B

Gemma 3 1B

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

Gemma 3 1B is an open-weights language model from Google, released March 2025. Its published size is 1B parameters and a 32k-token context window. Licence: gemma.

30.8±0.7#636 of 719 LLMsfrom 51 benchmark results across 13 sources

$ how it compares

The LLMs ranked either side of Gemma 3 1B, out of 828 with a score, then more from Google.

#modelscorewhereRAMreleased
731Qwen2 1.5B Instruct Alibaba31±1local~2.4GB2024-06
732Qwen-7B-Chat Alibaba31±1local~6.6GB2023-08
733Gemma 3 1B Google31±1local~2.2GB2025-03
734StarCoderBase 7B BigCode31±1local~6GB2023-07
735Llama 2 7B (base) Meta31±1local~5.5GB2023-07
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

agent tasks▃▁▃▆█▆▅▅▅▄▃▁ #300 of 300 · 2.4% of its score · 2 tests

Trails Falcon3 1B Instruct, Granite 4.0 350M, Qwen2.5 0.5B Instruct here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
BFCL v3 +116.3% ✎▃▁▁▁▁▃▅█▅▄▁▁ beats 0/114medium · avg 52%▇ 1.4%blunt · at its level · several runs
BFCL V4 (December 2025 archive)7.2% ✎▃▁▁▂██▅▅▆▂▂▂ beats 0/77hard · avg 32%▇ 1.0%typical
coding▁▁▁▃▆██▇▆▄▃▁ #437 of 458 · 10% of its score · 4 tests

Beats Qwen3.5 2B, Molmo 7B-D (0924), DeepSeek R1 Distill Qwen 1.5B here, though they rank higher overall. Trails Falcon3 1B Instruct, Granite 3.0 1B-A400M Instruct, Llama 3.2 1B Instruct here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
HumanEval lab41.5% ✎▁▁▁▁▁▃▅▅█▃▂▁ beats 58/190easy · avg 85%▇▇ 5.8%sharp · at its level
MBPP lab35.2% ✎▁▁▁▄▅█▄▅▃▁▂▁ beats 7/81easy · avg 72%▇ 2.7%at its level
LiveCodeBench (version unspecified) lab1.9% ✎▃···▁▁▃▅▅█▄▅ beats 0/37hard · avg 25%▇ 1.0%too hard for it
BIRD-SQL (dev) lab6.4% ✎▃·······▂▂█▆ beats 0/11medium · avg 44%▇ 0.7%typical
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #300 of 393 · 7.9% of its score · 4 tests

Beats Ministral 3 14B Reasoning 2512, Gemma 4 E4B, OpenThinker3 7B here, though they rank higher overall. Trails LFM2 350M here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
IFEval (aggregation unspecified) +162.9% ✎▄····█▄···▄█ beats 1/6easy · avg 82%▇ 3.2%at its level · several runs
IFEval mean of four measures63.2% ✎█·█··█·██·██ beats 1/6very easy · avg 87%▇ 2.4%at its level
IFEval prompt-level loose60.6% ✎▁▃▁▃▁▄▁▁█▃▃▁ beats 4/22easy · avg 80%▇ 1.3%at its level
AlpacaEval 2 length-controlled20.4% ✎▃▃·▃▁█▇▅·▁·▁ beats 4/22medium · avg 42%▇ 1.0%typical
math▁▂▃▅▆▅▄█▄▁▁▁ #352 of 398 · 31% of its score · 6 tests

Beats Yi 34B Chat, Llama 3 8B Instruct, Command R (03-2024) here, though they rank higher overall.

testresultvs the others in its rundifficultycountswhy
GSM8K +335.0% ✎▃▃··▃▂▅█▆▃▂▃ beats 0/22very easy · avg 93%▇▇▇▇▇▇▇▇ 19%sharp · at its level · several runs
MATH (subset unspecified) +140.3% ✎▃▃█▂▂▆·····▁ beats 8/22easy · avg 83%▇▇ 5.3%at its level · several runs
MGSM43.6% ✎█···█▄··▄··▄ beats 3/6easy · avg 77%▇▇ 4.4%at its level
MATH-50045.2% ✎▄··█▄·▄···▄▄ beats 2/6easy · avg 80%▇ 1.6%at its level
AIME 20251.0% ✎▄█··▄···▄··█ beats 2/6hard · avg 33%▇ 0.6%sharp · too hard for it · noisy
OTIS mock AIME 2024–20251.1% ✎▃▃▄▃▃▃▇▇█▃▂▅ beats 10/193very hard · avg 8.5%▇ 0.0%sharp · too hard for it · noisy
other tests22 not in a task · 48% of its score
testresultvs the others in its rundifficultycountswhy
MMLU +238.9% ✎▁▃·▁·▁▁█·▃▃▅ beats 1/22easy · avg 77%▇▇▇▇▇ 12%at its level · several runs
BIG-Bench Hard +139.4% ✎▁▄·▁▁·▅█·▄▄▁ beats 1/22easy · avg 81%▇▇▇▇ 9.0%sharp · several runs
MIMIC-CDM full information (4 pathologies)21.9% ✎▄·······▄█·▄ beats 0/4easy · avg 73%▇ 2.7%at its level
IFBench (aggregation unspecified) +120.5% ✎▄·█·█······█ beats 1/6medium · avg 40%▇ 2.7%several runs
MMLU-Pro +114.0% ✎█···█·██·███ beats 0/6medium · avg 54%▇ 2.3%small run · several runs
GPQA Diamond +219.9% ✎▃▁▁▁▂▄▆▄▅▇█▅ beats 3/208hard · avg 38%▇ 2.1%sharp · too hard for it · small run · several runs
MIMIC-CDM clinical decision making (4 pathologies)21.4% ✎█······█··██ beats 0/3medium · avg 62%▇ 2.1%at its level
Global-MMLU-Lite lab34.2% ✎▃···▂·▂▂▂·██ beats 0/12medium · avg 65%▇ 2.0%typical
14 more tests
testresultvs the others in its rundifficultycountswhy
FACTS Grounding lab36.4% ✎▄········▄·█ beats 0/3easy · avg 73%▇ 1.9%at its level
Open Persian LLM Leaderboard v2 average (21 tasks)37.5% ✎▂▄▃▃█▄▃▆▃▅▂▂ beats 12/81medium · avg 58%▇ 1.8%blunt · at its level
MMMLU34.4% ✎▄···▄·▄▄··▄█ beats 1/6medium · avg 59%▇ 1.7%typical
BRIDGE clinical text, zero-shot average (87 tasks)15.7% ✎▁▁▁▃▁▂▂▃▄▆█▃ beats 5/97hard · avg 32%▇ 1.5%blunt
TruthfulQA MC243.8% ✎▃▃▂·▃▃·▅█▂▅▂ beats 1/21medium · avg 63%▇ 1.0%blunt · at its level
GPQA (subset unspecified)21.1% ✎█······▄▄█·▄ beats 0/6hard · avg 39%▇ 0.8%sharp · too hard for it · noisy
Big Bench Extra Hard micro avg lab7.2% ✎▄▄·█▄·▄·▄·▄▄ beats 0/8hard · avg 19%▇ 0.8%too hard for it · small run
Open PT-LLM Leaderboard average (9 Portuguese tasks)54.2% ✎▃▁▁▁▁▁▁▃▂█▅▁ beats 74/276medium · avg 73%▇ 0.7%noisy
SimpleQA lab2.2% ✎▄▄▄▆█▆███·▂█ beats 0/32very hard · avg 7.4%▇ 0.7%too hard for it
EuroEval Life in the UK (English knowledge MC), test split, 5-shot, accuracy43.7% ✎▁▁▁▁▁▁▃▂▄█▄▃ beats 15/146easy · avg 83%▇ 0.7%at its level · noisy
PopQA9.6% ✎▃▁▁▃▃▃▁█▁▃·▄ beats 0/21hard · avg 32%▇ 0.7%blunt
WMT24++ (chrF) lab35.9% ✎█······█··██ beats 0/3medium · avg 53%▇ 0.7%blunt · at its level
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)94.7% ✎▁▁▂▃▄▄█▆▃▂▂▁ beats 58/140very easy · avg 94%▇ 0.6%blunt · noisy
Chess Puzzles (Epoch)0.0% ✎█▁▁▁▁▂▁▄▂▂▁▁ beats 29/135very hard · avg 5.9%▇ 0.0%sharp · too hard for it · 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 3 1B · needs ~0.8 GB · from huggingface.co/ggml-org
    chat in the terminal
    llama cli -hf ggml-org/gemma-3-1b-it-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf ggml-org/gemma-3-1b-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 3 1B · needs ~0.8 GB · from huggingface.co/ggml-org

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/ggml-org/gemma-3-1b-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 3 1B · needs ~1.1 GB · from ollama.com
    chat in the terminal
    ollama run gemma3:1b

MLX (Mac)

version

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run Gemma 3 1B · needs ~0.7 GB, Mac only · from huggingface.co/mlx-community
    chat in the terminal
    mlx_lm.chat --model mlx-community/gemma-3-1b-it-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model mlx-community/gemma-3-1b-it-4bit

Official weights and documentation: huggingface.co/google ↗

$ how the score was worked out · Google DeepMind Gemma 3 270M… 35%, 2 OLMo 2 Furious (Team OLMo,… 19%, LiquidAI/LFM2-1.2B model card 16% and 10 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
Google DeepMind Gemma 3 270M IT model card · lab-reported13▇▇▇▇▇▇▇▇▇▇ 35%CC-BY-4.0
2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models8▇▇▇▇▇▇ 19%CC BY 4.0
LiquidAI/LFM2-1.2B model card7▇▇▇▇▇ 16%LFM Open License v1.0
LiquidAI/LFM2.5-1.2B-Thinking model card8▇▇▇▇ 13%LFM Open License v1.0
MIMIC Clinical Decision Making leaderboard2▇ 5%Apache-2.0
Gemma Team — Gemma 3 Technical Report · lab-reported3▇ 4%CC BY 4.0
Berkeley Function Calling Leaderboard result archive2▇ 2%Apache-2.0
Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)1▇ 2%Apache-2.0
BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)1▇ 2%MIT
Open Portuguese LLM Leaderboard (archived)1▇ 1%Apache-2.0
EuroEval leaderboards (results collection), English generative leaderboard1▇ 1%MIT
Epoch AI — internal evaluations3▇ 1%CC BY 4.0
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)1▇ 1%Apache-2.0
prior · pull toward the middle▇ 0%
weighted average= 30.8

$ 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