[whatmodel.sh](../../README.md) / Qwen2.5 32B Instruct
# Qwen2.5 32B Instruct
Alibaba · released 2024-09 · 32.8B params · apache-2.0
**score 55.6** ±0.6 · #266 of 732 · from 35 benchmark results across 17 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Qwen2.5 32B Instruct among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃▆●▄▅▅▄▄▂▁ #210 of 313* 0.8% 1
coding ▁▁▁▃▅▇█●▅▄▃▁ #200 of 469 20% 6
chat & writing ▁▁▁▃▆▄▆●▆▇▇▃ #233 of 402 5.7% 3
math ▁▁▃▅█●▄█▄▂▁▁ #246 of 431 18% 6
agent tasks · #210 of 313 · 0.8% 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
BFCL v3 62.8% ▁▁▁▁▁▃▅█●▄▁▁ beats 87/114 medium · avg 51% 0.8% blunt · at its level
coding · #200 of 469 · 20% of its score
Beats [Gemini 2.5 Flash](gemini-2-5-flash.md), [Qwen3 30B-A3B Instruct 2507](qwen3-30b-a3b-instruct-2507.md), [Qwen3.5 4B](qwen3-5-4b.md) here, though they rank higher overall. Trails [QwQ-32B-Preview](qwq-32b-preview.md), [DeepSeek-Coder-V2 Instruct (0724)](deepseek-coder-v2-0724.md), [Ministral 3 8B Reasoning 2512](ministral-3-8b-reasoning.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LiveCodeBench official (Jan 1–Nov… 40.1% ▁▁·▁▂▄▆█●▂·▂ beats 24/31 hard · avg 23% 8.6% sharp · at its level
CRUXEval (ZeroEval, 800 problems) 69.9% ▁▂▆▃▆●▆▃▂·▁▁ beats 28/47 medium · avg 57% 7.9% sharp · at its level
CodeReviewQA: change localisation… 67.6% █▂▁▁▂▂▂▃▅▅▅● beats 46/47 hard · avg 33% 1.4% sharp · at its level · noisy
CodeReviewQA: automated code refi… 45.0% ▁▁▁▁▁▁▂▂▃█▅● beats 42/47 hard · avg 32% 1.0% at its level
CodeReviewQA: solution identifica… 89.5% ▃··▁▂▂▃▂█▃▄● beats 46/47 medium · avg 51% 1.0% sharp · at its level · noisy
CodeReviewQA: change type recogni… 71.0% ▁▁▁······▁▃● beats 20/47 easy · avg 67% 0.0% blunt · at its level · noisy
chat & writing · #233 of 402 · 5.7% of its score
Beats [Mistral Small 4 (119B-A6.5B)](mistral-small-4.md), [Olmo 3.1 32B Think](olmo-3-1-32b-think.md), [Qwen2.5 Coder 32B Instruct](qwen2-5-coder-32b.md) here, though they rank higher overall. Trails [Flan-T5 XXL](flan-t5-xxl.md), [LFM2.5 1.2B Instruct](lfm2-5-1-2b-instruct.md), [Gemma 3 4B](gemma-3-4b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arabic IFEval prompt-level (Incep… 60.4% ▁·▁▁▂▁▄▄▆▄█● beats 34/38 medium · avg 46% 2.8% sharp · at its level
AlpacaEval 2 length-controlled 39.1% ▃▃·▁▁█●▅·▁·▁ beats 14/22 medium · avg 39% 2.4% sharp · at its level
IFEval prompt-level loose 82.4% ▁▃▁▃▁▄▁▁●▃▃▁ beats 12/22 easy · avg 78% 0.6% typical
math · #246 of 431 · 18% of its score
Beats [GPT-4o (2024-08-06)](gpt-4o-2024-08-06.md), [Command A (03-2025)](command-a.md), [GPT-4o (2024-05-13)](gpt-4o-2024-05-13.md) here, though they rank higher overall. Trails [Gemma 7B (base)](gemma-7b-base.md), [CodeGemma 7B (v1.0)](codegemma-7b.md), [Phi-2 (2.7B)](phi-2.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
MATH level 5 56.1% ▂▂▅▆▄●▆▃▅▂▃▄ beats 48/92 hard · avg 28% 8.0% sharp · at its level
MATH Level 5 (ZeroEval, 721 probl… 59.5% ▁▁▁▆█▂▇▇●··▂ beats 33/38 hard · avg 30% 4.9% sharp · at its level
OTIS mock AIME 2024–2025 7.4% ▃▃▄●▃▃▇▇█▃▂▅ beats 42/197 very hard · avg 7.2% 2.0% sharp · noisy
MATH (subset unspecified) 77.9% ▃▃█▂▂●·····▁ beats 21/22 medium · avg 58% 1.7% at its level · several runs
GSM8K 87.5% ▂▃··▃▂▅█●▃▂▃ beats 15/22 very easy · avg 88% 1.0% sharp · maxed out
OMEGA (math generalisation) 19.2% ▅···●▃···▃▃█ beats 2/8 hard · avg 24% 0.9% at its level
other tests (18)
test result vs the others in its run difficulty counts why
MMLU-Redux (ZeroEval, 2,778 quest… 83.7% ▁·▁▁▄▄▄▅●▄▂▁ beats 41/53 easy · avg 75% 12% typical
ZebraLogic (ZeroEval, 1,000 puzzl… 26.1% ▁▁▃▆█●▁▁·▁▁▁ beats 40/57 hard · avg 15% 7.2% sharp · at its level
Medmarks-V mean win rate (19 veri… 0.5003 ▁▄▃▃▇▄●▃▅▁▁▁ beats 16/30 avg 0.418 6.7% sharp
GPQA Diamond 46.1% ▁▁▁▁▂▄●▄▅▇█▅ beats 53/211 hard · avg 39% 5.5% sharp · at its level
MMLU 84.7% ▁▂·▁·▁▁█·▃●▅ beats 18/22 easy · avg 75% 4.0% typical
BIG-Bench Hard 82.3% ▁▄·▁▁·▅█·●▄▁ beats 18/22 medium · avg 72% 3.6% sharp
Medmarks-V weighted mean win rate… 50.0% ▂▃▅▂▆▃●▅█▅▃▆ beats 16/34 medium · avg 41% 3.0% blunt · at its level
CPTUB average (SpeakLeash, 0-5 ju… 75.0% ▁▂▂▂▃█▃▆●▄▂▃ beats 70/95 medium · avg 65% 2.3% at its level
BRIDGE clinical text, zero-shot a… 39.9% ▁▁▁▁▁▂▂▃▄▆●▃ beats 73/97 hard · avg 31% 2.2% blunt · at its level
Polish EQ-Bench v2 (SpeakLeash) 71.2% ▁··▁▁▂▂▆▅▆●▆ beats 56/70 medium · avg 62% 1.8% at its level
EvalPerf correctness (118 tasks, … 87.6% ▂·▂▂▂▆▅▆▆▆▃● beats 25/29 easy · avg 74% 1.6% typical
Open Persian LLM Leaderboard v2 a… 62.8% ▂▄▂▃█▄▃●▃▅▂▂ beats 57/81 medium · avg 57% 1.6% blunt · at its level
PopQA 26.1% ▁▁▁▃▃▃●█▁▃·▄ beats 9/21 hard · avg 31% 1.2% at its level
SurgiQ surgical MCQ (zero-shot, l… 64.9% ▂···▂▂·▃▂▃█● beats 15/17 medium · avg 53% 0.9% at its level
Open PT-LLM Leaderboard average (… 80.2% ▃▁▁▁▁▁▁▁▂█●▁ beats 261/276 medium · avg 72% 0.7% at its level · noisy
Chess Puzzles (Epoch) 0.0% ●▁▁▁▁▂▃▃▃▂▁▁ beats 40/137 very hard · avg 5.8% 0.3% sharp · too hard for it · noisy
Vectara hallucination leaderboard… 97.0% ▁▁▂▃▄▄█●▃▂▂▁ beats 104/140 very easy · avg 94% 0.2% blunt · maxed out · noisy
TruthfulQA MC2 70.6% ▃▃▂·▃▃·▅█▂●▂ beats 19/21 medium · avg 60% 0.1% blunt · at its level
source results share of the score
WildEval ZeroEval published results 4 32% ▇▇▇▇
Epoch AI 4 16% ▇▇
2 OLMo 2 Furious 8 14% ▇▇
Medmarks: A Comprehensive Open-Source L… 2 10% ▇
LiveCodeBench official website results 1 9% ▇
CodeReviewQA: The Code Review Comprehen… 4 3% ▇
Inception Arabic/Hindi leaderboards 1 3% ▇
SpeakLeash CPTUB 1 2% ▇
BRIDGE Medical Leaderboard 1 2% ▇
SpeakLeash Polish EQ-Bench leaderboard 1 2% ▇
EvalPlus / EvalPerf published evaluatio… 1 2% ▇
Open Persian LLM Leaderboard v2 1 2% ▇
allenai/Olmo-3.1-32B-Think model card 2 1% ▇
SurgiQ: A Large-Scale Multi-Domain Benc… 1 1% ▇
Berkeley Function Calling Leaderboard r… 1 1% ▇
Open Portuguese LLM Leaderboard 1 1% ▇
Vectara hallucination leaderboard 1 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 55.6
Licences: WildEval ZeroEval published results ([Apache-2.0](../../licenses/zeroeval-published-results-Apache-2.0.txt)) · Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · 2 OLMo 2 Furious ([CC BY 4.0](../../licenses/paper-2501.00656v3-CC-BY-4.0.txt)) · Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks ([CC BY 4.0](../../licenses/paper-2605.01417v1-CC-BY-4.0.txt)) · LiveCodeBench official website results ([CC-BY-SA-4.0](../../licenses/lcb-sharealike-NOTICE.txt)) · CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · Inception Arabic/Hindi leaderboards ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · SpeakLeash CPTUB ([MIT](../../licenses/hf-leaderboard-spaces-MIT.txt)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · SpeakLeash Polish EQ-Bench leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · EvalPlus / EvalPerf published evaluation runs ([Apache-2.0](../../licenses/evalplus-published-runs-Apache-2.0.txt)) · Open Persian LLM Leaderboard v2 ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · allenai/Olmo-3.1-32B-Think model card ([Apache-2.0](../../licenses/allenai--Olmo-3.1-32B-Think.txt)) · SurgiQ: A Large-Scale Multi-Domain Benchmark for Evaluating Surgical Understanding in Large Language Models ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-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 Qwen2.5 32B Instruct (Q4_K_M, needs ~19.9 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-GGUF/tree/a15e3cc10f8bbb2c0af6f8f1f34a32e3b060c09d))
```sh
llama cli -hf Qwen/Qwen2.5-32B-Instruct-GGUF:q4_k_m
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf Qwen/Qwen2.5-32B-Instruct-GGUF:q4_k_m
```
Other versions: Q8_0 (~34.8 GB) `llama cli -hf Qwen/Qwen2.5-32B-Instruct-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 Qwen2.5 32B Instruct (Q4_K_M, needs ~19.9 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-GGUF/tree/a15e3cc10f8bbb2c0af6f8f1f34a32e3b060c09d))
```sh
lms get https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-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/Qwen2.5-32B-Instruct-GGUF@q8_0 -y` · 4bit (~18.4 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/Qwen2.5-32B-Instruct-MLX-4bit -y` · 8bit (~34.8 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/Qwen2.5-32B-Instruct-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 Qwen2.5 32B Instruct (Q4_K_M, needs ~20.18 GB, from [ollama.com](https://ollama.com/library/qwen2.5/tags))
```sh
ollama run qwen2.5:32b
```
Other versions: Q8_0 (~39.86 GB) `ollama run qwen2.5:32b-instruct-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 Qwen2.5 32B Instruct (4bit, needs ~18.4 GB, Mac only, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/Qwen2.5-32B-Instruct-MLX-4bit/tree/b3f0bad02db4e90b046baa3ed4bad1040d21f04b))
```sh
mlx_lm.chat --model lmstudio-community/Qwen2.5-32B-Instruct-MLX-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model lmstudio-community/Qwen2.5-32B-Instruct-MLX-4bit
```
Other versions: 8bit (~34.8 GB, Mac only) `mlx_lm.chat --model lmstudio-community/Qwen2.5-32B-Instruct-MLX-8bit`
Model repository and documentation: [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct)
[interactive version](https://whatmodel.sh/model/qwen2-5-32b) · [JSON](https://whatmodel.sh/api?format=json&all=1)