[whatmodel.sh](../../README.md) / Qwen3 4B
# Qwen3 4B
Alibaba · released 2025-04 · 4.02B params · apache-2.0
**score 51.2** ±0.5 · #345 of 732 · from 81 benchmark results across 14 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Qwen3 4B among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃▆●▄▅▅▄▄▂▁ #225 of 313 16% 8
coding ▁▁▁▃▅●█▇▅▄▃▁ #305 of 469 7.5% 4
chat & writing ▁▁▁▃▆▄●█▆▇▇▃ #251 of 402 9.0% 5
math ▁▁▃▅█●▄█▄▂▁▁ #224 of 431 14% 5
agent tasks · #225 of 313 · 16% of its score
Beats [o1 (2024-12-17)](o1-2024-12-17.md), [o3-mini](o3-mini.md), [DeepSeek R1 (Jan '25)](deepseek-r1.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Jamba 1.5 Mini](jamba-1-5-mini.md), [Mellum2 12B-A2.5B Instruct](mellum2-12b-a2-5b-instruct.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
τ²-bench Telecom (maker) 17.5% ▁·●█▆▅▂▃▃▂▃▁ beats 11/89 very hard · avg 13% 5.0% sharp · small run · several runs
τ-bench Retail (maker) 33.9% ▁····●·▂▇█▂▆ beats 3/28 hard · avg 30% 2.6% at its level · small run · several runs
τ²-bench Retail (maker) 38.6% ▁····●▁▃▆█▃▁ beats 2/44 hard · avg 28% 2.3% at its level · small run · several runs
τ-bench Airline (maker) 32.0% ▁▁▁·●▄▁█▇▁▁▇ beats 3/28 hard · avg 28% 1.8% at its level · small run · several runs
BFCL v3 63.4% ▁▁▁▁▁▃▅█●▄▁▁ beats 91/114 medium · avg 51% 1.6% blunt · at its level · several runs
τ²-bench Airline (maker) 28.0% ▁●·▂▁▃▂▁█▅▂▃ beats 1/42 hard · avg 17% 1.6% small run · several runs
EEGAgentBench overall (tool-based… 29.9% ▂·●▁··▇▇█▂·▁ beats 2/25 hard · avg 28% 0.8% blunt · at its level
BFCL (version not stated) 95.5% ▄▆▄▂·█▆·▄··● beats 17/17 very easy · avg 88% 0.5% blunt · small run
coding · #305 of 469 · 7.5% of its score
Beats [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md), [Solar Open 100B (102B-A12B)](solar-open-100b.md), [Qwen3-VL 30B-A3B Instruct](qwen3-vl-30b-a3b-instruct.md) here, though they rank higher overall. Trails [DeepSeek LLM 67B Chat](deepseek-llm-67b-chat.md), [WizardLM 70B V1.0](wizardlm-70b-v1-0.md), [Yi 1.5 9B Chat](yi-1-5-9b-chat.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Aider polyglot (225 exercises, tw… 13.8% ▂▁▁▃●▃▄▆█▃▄▂ beats 13/71 very hard · avg 8.3% 3.3% sharp · small run
LiveCodeBench v6 (Feb–May 2025) (… 48.4% ▁▁▃●█▅▄▇▅▅▃█ beats 4/38 medium · avg 43% 2.4% at its level · small run · several runs
MultiPL-E (maker) 66.6% ▃··▄·▅▅●▄█▃▄ beats 15/31 easy · avg 65% 1.0% at its level · small run
LiveCodeBench v4 52.9% █·█········● beats 2/2 medium · avg 45% 0.8% at its level · small run
chat & writing · #251 of 402 · 9.0% of its score
Beats [Mistral Small 4 (119B-A6.5B)](mistral-small-4.md), [Qwen2.5 Coder 32B Instruct](qwen2-5-coder-32b.md), [Ministral 3 14B Reasoning 2512](ministral-3-14b-reasoning.md) here, though they rank higher overall. Trails [Flan-T5 XXL](flan-t5-xxl.md), [Granite 4.0 H 1B](granite-4-0-h-1b.md), [LFM2 2.6B](lfm2-2-6b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena-Hard v2 aggregate (maker) 13.7% ●▁··▄▃▁▃█▁▁▇ beats 0/22 hard · avg 20% 2.9% sharp · small run · several runs
IFEval (aggregation unspecified) … 81.9% ▁▂▄▂▂▄▆●█▅▂▁ beats 76/145 easy · avg 80% 2.2% small run · several runs
Arena-Hard v2.0 hard prompts (sty… 15.0% ▂·▃●▅·▃█▆▂▂▂ beats 4/22 very hard · avg 12% 1.9% sharp
IFEval mean of four measures 85.6% █··▄▄·▄····● beats 5/5 easy · avg 84% 1.4% typical
Arena-Hard v2.0 creative writing … 13.2% ▆●▂▆▂▂▃█▂▃·▂ beats 4/22 hard · avg 19% 0.5% sharp · noisy
math · #224 of 431 · 14% of its score
Beats [MiniMax-M2.1](minimax-m2-1.md), [DeepSeek V3 (Dec '24)](deepseek-v3.md), [GPT-4o (2024-08-06)](gpt-4o-2024-08-06.md) here, though they rank higher overall. Trails [Phi-2 (2.7B)](phi-2.md), [Granite 8B Code Base (4K)](granite-8b-code-base-4k.md), [Mixtral 8x7B v0.1 (base)](mixtral-8x7b-v0-1.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
AIME 2025 (maker) 65.6% ▁·▁▁▁▂▅▆●▇▅▁ beats 56/133 medium · avg 59% 7.2% sharp · at its level · small run · seve…
HMMT 2025 (edition unspecified) (… 42.1% ▃●▇▅▇▅▁·▁··▁ beats 5/32 hard · avg 34% 4.3% sharp · at its level · small run · seve…
PolyMATH (maker) 40.0% ▁···▃▁·●█▄▄▄ beats 4/23 hard · avg 35% 1.8% at its level · small run · several runs
MATH-500 85.6% ▃····▃·█···● beats 5/5 easy · avg 79% 0.9% typical
GSM8K loose substring accuracy 68.5% ●·█··███···█ beats 0/5 easy · avg 83% 0.2% blunt
other tests (30)
test result vs the others in its run difficulty counts why
GPQA Diamond 52.3% ▁▁▁▁▂▄▆●▅▇█▅ beats 74/211 medium · avg 52% 11% sharp · at its level · small run · seve…
MMLU-Pro 52.3% ███···██···● beats 5/5 medium · avg 42% 5.0% at its level · several runs
GPQA (subset unspecified) (maker) 55.9% ▂▁▂▂▄▃▆▃●█▃▃ beats 59/106 hard · avg 51% 4.6% sharp · at its level · small run · seve…
MMLU-Redux (maker) 83.7% ▁▁▃▁▂▃●▅▇▅█▄ beats 15/63 easy · avg 77% 2.8% at its level · small run · several runs
MMLU 72.3% ▄█···█·····● beats 5/5 medium · avg 73% 2.5% at its level
INCLUDE (maker) 61.8% ▂····▂▂▄●▄█▄ beats 13/36 medium · avg 55% 2.4% at its level · small run · several runs
BRIDGE clinical text, zero-shot a… 38.5% ▁▁▁▁▁▂▂▃▄▆●▃ beats 63/97 hard · avg 31% 1.8% blunt · at its level
IFBench (aggregation unspecified) 30.3% █··███··█··● beats 5/5 hard · avg 31% 1.8% at its level
SuperGPQA (maker) 42.7% ▂▁▂▂▅●▃▇▇▇█▃ beats 13/46 medium · avg 39% 1.7% at its level · small run · several runs
MMLU-ProX (maker) 61.0% ▁·▁▁▁●▁▃▃█▅▄ beats 4/36 medium · avg 54% 1.6% at its level · small run · several runs
WritingBench (maker) 73.5% ●·▁▂▁··▅█▂▂▁ beats 0/20 easy · avg 71% 1.5% at its level · small run · several runs
Creative Writing v3 (maker) 61.1% ●····▂▆··▆█▆ beats 0/14 easy · avg 66% 1.5% at its level · small run · several runs
LiveBench 2024-11-25 (maker) 63.6% ●··▂·▂·▂█▆▂▂ beats 1/13 medium · avg 62% 1.5% at its level · small run · several runs
Open Persian LLM Leaderboard v2 a… 50.5% ▂▄▂▃●▄▃▆▃▅▂▂ beats 33/81 medium · avg 57% 1.5% blunt · at its level
Multi-IF (aggregation unspecified… 66.3% ▁▁▁▁▁▁●▄█▄▁▂ beats 9/41 medium · avg 62% 1.4% at its level · small run · several runs
KINA (Knowledge Index of Noah's A… 21.5% ▁▁▂▁●▄█▆▄▅▂▃ beats 5/39 very hard · avg 21% 1.3% typical
MMMLU seven languages (Liquid) 60.7% ▄·█····█···● beats 5/5 medium · avg 59% 1.3% at its level
Multi-IF three-turn average 75.5% ▄··█▄··▄···● beats 5/5 easy · avg 70% 1.2% at its level
OJBench (maker) 16.1% ▁···▁▃·●▄▅██ beats 4/26 very hard · avg 13% 1.0% small run
Global-MMLU 73.3% ▃▅▃·●···▅·▅█ beats 4/11 medium · avg 70% 0.8% at its level · small run
Open PT-LLM Leaderboard average (… 76.2% ▃▁▁▁▁▁▁▁▂█●▁ beats 231/276 medium · avg 72% 0.7% at its level · noisy
MGSM six languages, loose (Liquid) 81.8% ▄·▄·█···●··▄ beats 4/5 easy · avg 81% 0.7% typical
MixEval-Hard 38.0% █····█·····● beats 2/2 medium · avg 39% 0.6% at its level · small run
GSM-Plus 56.2% █·····●██·██ beats 1/5 easy · avg 83% 0.6% blunt · small run · several runs
EuroEval HellaSwag (English), tes… 70.8% ▃·▁▁▂▁▂▅●▇▄▁ beats 53/100 medium · avg 67% 0.6% sharp · at its level · noisy
ZebraLogic (maker) 35.2% ●·▁▁▁▇·▄▄▁▄▁ beats 4/24 medium · avg 45% 0.5% at its level · noisy · small run
EuroEval Life in the UK (English … 82.4% ▁▁▁▁▁▁▁▂▄●▄▃ beats 78/146 easy · avg 82% 0.4% noisy
Vectara hallucination leaderboard… 97.3% ▁▁▂▃▄▄█●▃▂▂▁ beats 111/140 very easy · avg 94% 0.2% blunt · maxed out · noisy
Chess Puzzles (Epoch) 1.0% ●▁▁▁▁▂▃▃▃▂▁▁ beats 2/137 very hard · avg 5.8% 0.1% sharp · too hard for it · noisy
Vectara hallucination leaderboard… 94.3% ▁▂▁▂▅█▂▃●▃▁▁ beats 79/98 very easy · avg 90% 0.0% blunt
source results share of the score
Qwen/Qwen3-4B-Thinking-2507 model card … 23 31% ▇▇▇▇
Qwen/Qwen3-4B-Instruct-2507 model card … 25 29% ▇▇▇▇
LFM2 Technical Report 11 15% ▇▇
HuggingFaceTB/SmolLM3-3B model card 8 11% ▇
Epoch AI 2 4% ▇
Arena-Hard-Auto v2.0-Preview leaderboard 2 2% ▇
BRIDGE Medical Leaderboard 1 2% ▇
Open Persian LLM Leaderboard v2 1 1% ▇
Knowledge Index of Noah's Ark 1 1% ▇
EuroEval leaderboards 2 1% ▇
EEGAgentBench: Benchmarking LLM Agents … 1 1% ▇
Open Portuguese LLM Leaderboard 1 1% ▇
Berkeley Function Calling Leaderboard r… 1 1% ▇
Vectara hallucination leaderboard 2 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 51.2
Licences: Qwen/Qwen3-4B-Thinking-2507 model card ([apache-2.0](../../licenses/Qwen--Qwen3-4B-Thinking-2507.txt)) · Qwen/Qwen3-4B-Instruct-2507 model card ([apache-2.0](../../licenses/Qwen--Qwen3-4B-Instruct-2507.txt)) · LFM2 Technical Report ([CC BY 4.0](../../licenses/cards-gap-NOTICE.txt)) · HuggingFaceTB/SmolLM3-3B model card ([Apache-2.0](../../licenses/HuggingFaceTB--SmolLM3-3B.txt)) · Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Arena-Hard-Auto v2.0-Preview leaderboard ([Apache-2.0](../../licenses/arena-hard-auto-Apache-2.0.txt)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · Open Persian LLM Leaderboard v2 ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · Knowledge Index of Noah's Ark ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE)) · EEGAgentBench: Benchmarking LLM Agents on Short- and Long-Horizon EEG Analysis ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))
## 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 Qwen3 4B (Q4_K_M, needs ~2.5 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-4B-GGUF/tree/bc640142c66e1fdd12af0bd68f40445458f3869b))
```sh
llama cli -hf Qwen/Qwen3-4B-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf Qwen/Qwen3-4B-GGUF:Q4_K_M
```
Other versions: Q8_0 (~4.3 GB) `llama cli -hf Qwen/Qwen3-4B-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 Qwen3 4B (Q4_K_M, needs ~2.5 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-4B-GGUF/tree/bc640142c66e1fdd12af0bd68f40445458f3869b))
```sh
lms get https://huggingface.co/Qwen/Qwen3-4B-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 (~4.3 GB) `lms get https://huggingface.co/Qwen/Qwen3-4B-GGUF@Q8_0 -y` · 4bit (~2.1 GB, Mac only) `lms get https://huggingface.co/Qwen/Qwen3-4B-MLX-4bit -y` · 8bit (~4.1 GB, Mac only) `lms get https://huggingface.co/Qwen/Qwen3-4B-MLX-8bit -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 Qwen3 4B (Q4_K_M, needs ~2.91 GB, from [ollama.com](https://ollama.com/library/qwen3/tags))
```sh
ollama run qwen3:4b-q4_K_M
```
Other versions: Q8_0 (~5.32 GB) `ollama run qwen3:4b-q8_0`
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run Qwen3 4B (4bit, needs ~2.1 GB, Mac only, from [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-4B-MLX-4bit/tree/52a5ab34fa604bc8af6d3ce0cac0cab10b7eb495))
```sh
mlx_lm.chat --model Qwen/Qwen3-4B-MLX-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model Qwen/Qwen3-4B-MLX-4bit
```
Other versions: 8bit (~4.1 GB, Mac only) `mlx_lm.chat --model Qwen/Qwen3-4B-MLX-8bit`
Model repository and documentation: [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-4B)
[interactive version](https://whatmodel.sh/model/qwen3-4b) · [JSON](https://whatmodel.sh/api?format=json&all=1)