[whatmodel.sh](../../README.md) / QwQ-32B
# QwQ-32B
Alibaba · released 2025-03 · 32.76B params · apache-2.0
**score 59.4** ±0.8 · #212 of 732 · from 27 benchmark results across 13 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks QwQ-32B among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃▆●▄▅▅▄▄▂▁ #235 of 313* 3.0% 2
coding ▁▁▁▃▅▇█●▅▄▃▁ #210 of 469 28% 7
chat & writing ▁▁▁▃▆▄▆█●▇▇▃ #172 of 402 21% 5
math ▁▁▃▅█▆▄●▄▂▁▁ #152 of 431 8.3% 3
agent tasks · #235 of 313 · 3.0% of its score
Beats [o1 (2024-12-17)](o1-2024-12-17.md), [o3-mini](o3-mini.md), [Llama 3.1 Nemotron Ultra 253B v1](llama-3-1-nemotron-ultra-253b.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Jamba 1.5 Mini](jamba-1-5-mini.md), [LFM2.5 1.2B Thinking](lfm2-5-1-2b-thinking.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
BFCL V4 (2025-11-03 archive, mode… 36.9% ▃··▅●▃·▅·▃·█ beats 3/10 hard · avg 33% 1.5% at its level
BFCL v3 58.3% ▁▁▁▁▁▃▅█●▄▁▁ beats 75/114 medium · avg 51% 1.4% blunt · at its level
coding · #210 of 469 · 28% of its score
Beats [Qwen3.5 4B](qwen3-5-4b.md), [Falcon-H1R 7B](falcon-h1r-7b.md), [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md) here, though they rank higher overall. Trails [QwQ-32B-Preview](qwq-32b-preview.md), [Ministral 3 3B Reasoning 2512](ministral-3-3b-reasoning.md), [DeepSeek-Coder-V2 Instruct (0724)](deepseek-coder-v2-0724.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Aider polyglot (225 exercises, tw… 20.9% ▂·▂▃●▂▃█▆▄▂▂ beats 11/49 very hard · avg 7.6% 16% sharp · at its level
WebApp1K-Duo-React pass@1 (onekq-… 61.6% ▁··▁·▁▁▁▆●▇█ beats 16/45 hard · avg 29% 4.0% at its level
CodeReviewQA: change localisation… 63.8% █▂▁▁▂▂▂▃▅▅▅● beats 42/47 hard · avg 33% 2.3% sharp · at its level · noisy
WebApp1K-React pass@1 (onekq-ai) 87.0% ▃▁▁█▁▄▅▃▄●▁▄ beats 23/31 medium · avg 40% 2.3% sharp · noisy
CodeReviewQA: automated code refi… 32.9% ▁▁▁▁▁▁▂▂▃●▅▄ beats 30/47 hard · avg 32% 2.0% at its level
CodeReviewQA: solution identifica… 77.2% ▃··▁▂▂▃▂█▃●▂ beats 41/47 medium · avg 51% 1.4% sharp · noisy
CodeReviewQA: change type recogni… 60.9% ▁▁▁······▁●█ beats 10/47 easy · avg 67% 0.1% blunt · at its level · noisy
chat & writing · #172 of 402 · 21% of its score
Beats [MiniMax-M2](minimax-m2.md), [MiniCPM5 2B](minicpm5-2b.md), [Granite 4.2 30B](granite-4-2-30b.md) here, though they rank higher overall. Trails [LFM2.5 1.2B Instruct](lfm2-5-1-2b-instruct.md), [LFM2.5 1.2B Thinking](lfm2-5-1-2b-thinking.md), [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena-Hard v2.0 hard prompts (sty… 43.5% ▂·▃▅▅·▃●▆▂▂▂ beats 11/22 very hard · avg 12% 7.7% sharp · at its level
LMArena (text) 1335.6 ▁▁▁▂▃▃▃●▆▇█▃ beats 172/385 avg 1260 6.7% blunt
AlpacaEval 2 length-controlled 82.4% ▃▃·▁▁█▇▅·▁·● beats 22/22 medium · avg 39% 4.1% sharp · at its level
Arena-Hard v2.0 creative writing … 60.9% ▆▂▂▆▂▂▃●▂▃·▂ beats 16/22 hard · avg 19% 1.7% sharp · at its level · noisy
IFEval prompt-level loose 85.8% ▁▃▁▃▁▄▁▁█●▃▁ beats 19/22 easy · avg 78% 0.9% typical
math · #152 of 431 · 8.3% of its score
Beats [MiMo-V2.5 (310B-A15B)](mimo-v2-5.md), [MiniMax-M2.7](minimax-m2-7.md), [MiniMax-M2.1](minimax-m2-1.md) here, though they rank higher overall. Trails [Nemotron 3 Nano 30B-A3B](nemotron-3-nano-30b-a3b.md), [Qwen3-VL 8B Thinking](qwen3-vl-8b-thinking.md), [Olmo 3 32B Think](olmo-3-32b-think.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
OTIS mock AIME 2024–2025 59.2% ▃▃▄▃▃▃●▇█▃▂▅ beats 89/197 very hard · avg 7.2% 5.3% sharp · at its level · noisy
MATH (subset unspecified) 98.1% ▃▃█▂▂▆·····● beats 22/22 medium · avg 58% 1.8% typical
GSM8K 95.5% ▂▃··▃▂▅█▆▃▂● beats 22/22 very easy · avg 88% 1.2% sharp · maxed out
other tests (10)
test result vs the others in its run difficulty counts why
GPQA Diamond 65.3% ▁▁▁▁▂▄▆▄●▇█▅ beats 94/211 hard · avg 39% 11% sharp · at its level
MMLU 88.4% ▁▂·▁·▁▁█·▃▃● beats 22/22 easy · avg 75% 5.7% typical
BIG-Bench Hard 89.6% ▁▄·▁▁·▅█·▄▄● beats 22/22 medium · avg 72% 5.1% sharp
BRIDGE clinical text, zero-shot a… 39.4% ▁▁▁▁▁▂▂▃▄▆●▃ beats 67/97 hard · avg 31% 4.2% blunt · at its level
EQ-Bench 3 (emotional intelligenc… 1197.5 ▁·▂▁▄▅●▃█▄▁▁ beats 30/62 avg 1034 4.1% typical
ARC-AGI-1 public evaluation set (… 11.3% ▁·●▃▁·█▇▁·▇▃ beats 3/27 very hard · avg 4.0% 3.6% sharp · noisy
Open Persian LLM Leaderboard v2 a… 60.6% ▂▄▂▃█▄▃●▃▅▂▂ beats 52/81 medium · avg 57% 2.7% blunt · at its level
Chess Puzzles (Epoch) 5.0% ●▁▁▁▁▂▃▃▃▂▁▁ beats 0/137 very hard · avg 5.8% 1.1% sharp · too hard for it · noisy
EuroEval HellaSwag (English), tes… 85.3% ▃·▁▁▂▁▂▅█●▄▁ beats 84/100 medium · avg 67% 0.9% sharp · noisy
EuroEval Life in the UK (English … 93.2% ▁▁▁▁▁▁▁▂▄█▄● beats 129/146 easy · avg 82% 0.5% noisy
source results share of the score
2 OLMo 2 Furious 6 19% ▇▇
Epoch AI 3 18% ▇▇
Aider polyglot benchmark runs 1 16% ▇▇
Arena-Hard-Auto v2.0-Preview leaderboard 2 9% ▇
LMArena leaderboard dataset 1 7% ▇
WebApp1K models leaderboard 2 6% ▇
CodeReviewQA: The Code Review Comprehen… 4 6% ▇
BRIDGE Medical Leaderboard 1 4% ▇
EQ-Bench 3 canonical leaderboard results 1 4% ▇
ARC Prize Foundation model_baseline out… 1 4% ▇
Berkeley Function Calling Leaderboard r… 2 3% ▇
Open Persian LLM Leaderboard v2 1 3% ▇
EuroEval leaderboards 2 1% ▇
prior: pull toward the middle 0% ▇
weighted average = 59.4
Licences: 2 OLMo 2 Furious ([CC BY 4.0](../../licenses/paper-2501.00656v3-CC-BY-4.0.txt)) · Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Aider polyglot benchmark runs ([Apache-2.0](../../licenses/aider-polyglot-runs-Apache-2.0.txt)) · Arena-Hard-Auto v2.0-Preview leaderboard ([Apache-2.0](../../licenses/arena-hard-auto-Apache-2.0.txt)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · WebApp1K models leaderboard ([MIT](../../licenses/webapp1k-leaderboard-MIT.txt)) · CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · EQ-Bench 3 canonical leaderboard results ([MIT](../../licenses/eqbench3-MIT.txt)) · ARC Prize Foundation model\_baseline outputs on ARC-AGI-1/2 public evaluation sets ([MIT](../../licenses/arcprize-model-baseline-outputs-MIT.txt)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · Open Persian LLM Leaderboard v2 ([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 QwQ-32B (Q4_K_M, needs ~19.9 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/QwQ-32B-GGUF/tree/8728e66249190b78dee8404869827328527f6b3b))
```sh
llama cli -hf Qwen/QwQ-32B-GGUF:q4_k_m
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf Qwen/QwQ-32B-GGUF:q4_k_m
```
Other versions: Q8_0 (~34.8 GB) `llama cli -hf Qwen/QwQ-32B-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 QwQ-32B (Q4_K_M, needs ~19.9 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/QwQ-32B-GGUF/tree/8728e66249190b78dee8404869827328527f6b3b))
```sh
lms get https://huggingface.co/Qwen/QwQ-32B-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 (~34.8 GB) `lms get https://huggingface.co/Qwen/QwQ-32B-GGUF@q8_0 -y` · 4bit (~18.4 GB, Mac only) `lms get https://huggingface.co/mlx-community/QwQ-32B-4bit -y` · 8bit (~34.8 GB, Mac only) `lms get https://huggingface.co/mlx-community/QwQ-32B-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 QwQ-32B (Q4_K_M, needs ~20.16 GB, from [ollama.com](https://ollama.com/library/qwq/tags))
```sh
ollama run qwq:32b
```
Other versions: Q8_0 (~39.81 GB) `ollama run qwq:32b-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 QwQ-32B (4bit, needs ~18.4 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/QwQ-32B-4bit/tree/0e98026e38d9957ffdabff95384fd9ee096bb10d))
```sh
mlx_lm.chat --model mlx-community/QwQ-32B-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model mlx-community/QwQ-32B-4bit
```
Other versions: 8bit (~34.8 GB, Mac only) `mlx_lm.chat --model mlx-community/QwQ-32B-8bit`
Model repository and documentation: [huggingface.co/Qwen](https://huggingface.co/Qwen/QwQ-32B)
[interactive version](https://whatmodel.sh/model/qwq-32b) · [JSON](https://whatmodel.sh/api?format=json&all=1)