whatmodel.sh / Qwen3-VL 32B Instruct

Qwen3-VL 32B Instruct

+ contribute

Qwen3-VL 32B Instruct is an open-weights language model from Alibaba, released October 2025. Its published size is 33.4B parameters and a 256k-token context window. Licence: apache-2.0.

Catalogue identity notes

Maker repo Qwen/Qwen3-VL-32B-Instruct @0cfaf48 (licence tag apache-2.0, created 2025-10-19); the separate Instruct checkpoint, not Qwen3-VL-32B-Thinking. Scored from the Qwen3-VL Technical Report (arXiv 2511.21631v2, Tables 3/4 and 7-10: column 'Qwen3-VL 32B instruct'), whose instruct and thinking columns are the two published checkpoints. Released 2025-10: month the maker's Hugging Face repo was created.

Model information sources

Release date ↗ · Weights licence ↗

33,357,390,064 parameters in the Hugging Face safetensors metadata of Qwen/Qwen3-VL-32B-Instruct @0cfaf48 (vision encoder included); dense, so active = total.

60.1±0.8#194 of 719 LLMsfrom 25 benchmark results across 1 source

$ how it compares

The LLMs ranked either side of Qwen3-VL 32B Instruct, out of 828 with a score, then more from Alibaba.

#	model	score	where	RAM	released
230	DeepSeek V3 0324 DeepSeek	60±1	local	~108GB	2025-03
231	Qwen3 30B-A3B Instruct 2507 Alibaba	60±1	api $0.15/M +local	~20GB	2025-07
232	Qwen3-VL 32B Instruct Alibaba	60±1	api $0.28/M +local	~22GB	2025-10
233	Falcon-H1R 7B TII	60±1	local	~6GB	2026-01
234	Gemini 2.0 Flash (001) Google	60±1	comparison only		2025-02
	more from Alibaba
18	Qwen3.8 Max (2.4T-A95B) Alibaba	78±1	api $3.00/M +local	~380GB	2026-08
34	Qwen3.8 Max (0902) Alibaba	76±1	api $3.00/M		2026-09
37	Qwen3.8-Flash-Next Alibaba	75±1	local	~59GB	2026-08

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #160 of 300 * · 2.1% of its score · 1 test

Beats o1 (2024-12-17), o3-mini, DeepSeek R1 0528 here, though they rank higher overall. Trails SmolLM3 3B, Jamba 1.5 Mini, Jamba 1.5 Large here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
BFCL v3 lab	70.2% ✎	█▄··▄█·█▄█▄█ beats 11/13	medium · avg 59%	▇ 2.1%	blunt · at its level

coding▁▁▁▃▆██▇▆▄▃▁ #295 of 458 * · 4.7% of its score · 1 test

Trails DeepSeek LLM 67B Chat, WizardLM 70B V1.0, Yi 1.5 9B Chat here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
LiveCodeBench v6 (window unspecified) lab	43.8% ✎	▃▅█▅▅▃▅···▃█ beats 7/16	medium · avg 35%	▇▇ 4.7%	at its level

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #179 of 393 * · 2.7% of its score · 1 test

Beats Seed-OSS 36B Instruct, Nemotron-Cascade-14B-Thinking here, though they rank higher overall. Trails LFM2.5 1.2B Instruct, LFM2.5 1.2B Thinking, Gemma 3 12B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
IFEval (aggregation unspecified) lab	84.7% ✎	▃··▃···██▅▃▃ beats 8/11	easy · avg 77%	▇ 2.7%	typical

math▁▂▃▅▆▅▄█▄▁▁▁ #159 of 398 * · 10% of its score · 2 tests

Beats MiMo-V2.5 (310B-A15B), MiniMax-M2.7, MiniMax-M2.1 here, though they rank higher overall. Trails Olmo 3 7B Think, Qwen3-VL 4B Thinking, Granite 4.2 3B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
MathVista (testmini) lab	83.8% ✎	▃··▃▅·▃█▅▅▅█ beats 14/16	medium · avg 69%	▇▇ 5.2%	typical
AIME 2025 lab	66.2% ✎	▃·▃▃█··▃▅▃·█ beats 6/12	medium · avg 45%	▇▇ 4.4%	sharp · at its level

vision & documents▁▁▃▂▁▃▄▃▆▇█▄ #86 of 147 · 49% of its score · 13 tests

Beats DeepSeek V4 Flash Vision (Exp), GLM-4.6V, GPT-5 nano here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
CharXiv (reasoning questions) lab	62.8% ✎	▄▂▄▆▄▄▆··▄█▂ beats 12/21	medium · avg 39%	▇▇▇▇ 10%	at its level
MMMU (val) lab	76.0% ✎	▂···▆▂▂▂▄▄▄█ beats 13/16	medium · avg 61%	▇▇▇ 8.0%	at its level
CharXiv descriptive questions lab	90.5% ✎	▄··▂▂▂·▂▄▄▂█ beats 13/14	easy · avg 74%	▇▇ 5.6%	typical
MMStar lab	77.7% ✎	▂··▄▂▂·▄█▂▄▆ beats 14/16	medium · avg 65%	▇▇ 5.3%	at its level
MathVista (testmini) lab	83.8% ✎	▃··▃▅·▃█▅▅▅█ beats 14/16	medium · avg 69%	▇▇ 5.2%	typical
MMMU Pro lab	65.3% ✎	▃·█·▅▃▃▃▅▅▃█ beats 14/16	medium · avg 46%	▇▇ 4.9%	at its level
HallusionBench lab	63.8% ✎	▃·▃·▅▃▅▅▃▅█▅ beats 12/16	medium · avg 55%	▇ 2.2%	blunt · at its level
ChartQA lab	88.5% ✎	▂··▂·▂▃·▅▂▆█ beats 9/17	easy · avg 85%	▇ 1.5%	blunt

5 more tests

test	result	vs the others in its run	difficulty	counts	why
RealWorldQA lab	79.0% ✎	▅▃·▃▅▅·█·▃▅█ beats 15/16	easy · avg 70%	▇ 1.5%	blunt
ScreenSpot-Pro lab	57.9% ✎	▂····▂▄··▂█▆ beats 7/11	medium · avg 46%	▇ 1.4%	blunt · at its level
DocVQA lab	96.9% ✎	▃··▅▃▅▃█▃█▅▅ beats 16/17	very easy · avg 94%	▇ 1.3%	blunt · maxed out
CC-OCR lab	80.3% ✎	▃··▃▅·▅▅·█▅▃ beats 11/13	easy · avg 74%	▇ 1.0%	blunt
OCRBench lab	89.5% ✎	▅█·▃▃█▃▅▅▃·▃ beats 13/16	easy · avg 84%	▇ 0.5%	blunt

other tests8 not in a task · 37% of its score

test	result	vs the others in its run	difficulty	counts	why
GPQA (subset unspecified) lab	68.9% ✎	▂····▂·▂▂█▆█ beats 8/14	hard · avg 49%	▇▇▇▇ 9.4%	sharp · at its level
MMLU-Pro lab	78.6% ✎	▄·▄·▄▄▄▄█▄██ beats 10/12	medium · avg 63%	▇▇▇ 6.6%	typical
MMLU-Redux lab	89.8% ✎	▃··▃·▃·█▃▅▃▅ beats 9/11	easy · avg 79%	▇▇ 5.1%	typical
SuperGPQA lab	54.6% ✎	▄··▄·▄▄▄█▄██ beats 9/11	medium · avg 37%	▇▇ 4.9%	at its level
AndroidWorld lab	57.3% ✎	▃··▃▅▃▃▃▃▃·█ beats 8/11	medium · avg 43%	▇ 3.4%	blunt · at its level
LVBench lab	63.8% ✎	▅··▃·▅▃▃·█·▃ beats 9/10	hard · avg 51%	▇ 3.4%	at its level
Video-MME (without subtitles) lab	76.6% ✎	▃··█··▅█▃▃▅▃ beats 11/13	medium · avg 66%	▇ 3.0%	blunt · at its level
MVBench lab	72.8% ✎	▂▂▂·▂█··▆▄·▂ beats 10/13	medium · avg 67%	▇ 1.3%	blunt · at its level

$ where to run it

Through an API

api · Alibaba Cloud Model Studio · $0.28/M · $0.16 in / $0.64 out per 1M tokens · 256k context · www.alibabacloud.com ↗ · price sourceUSD per 1M tokens, Alibaba Cloud Model Studio, Singapore (International deployment scope), API model qwen3-vl-32b-instruct (Non-Thinking mode only): input $0.16, output $0.64. Chinese-mainland/Global (US) deployments priced differently and not used, following the International convention of other Alibaba rows; free quota, batch and context-cache discounts excluded. Observed 2026-10-09.

llama.cpp · recommended

version

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 ↗
Run Qwen3-VL 32B Instruct · needs ~19.8 GB · from huggingface.co/Qwen (official)
chat in the terminal
llama cli -hf Qwen/Qwen3-VL-32B-Instruct-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf Qwen/Qwen3-VL-32B-Instruct-GGUF:Q4_K_M

LM Studio

version

Install LM Studio · Open the app once before using the lms command.
download it (macOS / Windows / Linux) ↗
Run Qwen3-VL 32B Instruct · needs ~19.8 GB · from huggingface.co/Qwen (official)
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/Qwen/Qwen3-VL-32B-Instruct-GGUF@Q4_K_M -y
Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

Install Ollama
download it (macOS / Windows) ↗
Linux
curl -fsSL https://ollama.com/install.sh | sh
Run Qwen3-VL 32B Instruct · needs ~21 GB · from ollama.com
chat in the terminal
ollama run qwen3-vl:32b-instruct

MLX (Mac)

version

Install MLX (Mac) · Apple silicon Macs only, text only.
with uv
uv tool install mlx-lm
or
pip install mlx-lm
Run Qwen3-VL 32B Instruct · needs ~19.6 GB, Mac only · from huggingface.co/lmstudio-community
chat in the terminal
mlx_lm.chat --model lmstudio-community/Qwen3-VL-32B-Instruct-MLX-4bit
Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model lmstudio-community/Qwen3-VL-32B-Instruct-MLX-4bit

Official weights and documentation: huggingface.co/Qwen ↗

$ how the score was worked out · Qwen (Alibaba) vendor-reporte… 100%

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

source	results	share of the score	licence
Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card) · lab-reported	25	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 100%	Factual vendor-reported score (not licensed content)
prior · pull toward the middle		▇ 0%
weighted average		= 60.1

$ improve this page

Update model information Add a place to run it Add a test result Propose a benchmark Correct a benchmark score Edit 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 2026-10-11 02:20 UTC · sources and licences · json