whatmodel.sh / Qwen2-VL 2B Instruct

Qwen2-VL 2B Instruct

+ contribute

Qwen2-VL 2B Instruct is an open-weights language model from Alibaba, released August 2024. Its published size is 2.21B parameters and a 33k-token context window. Licence: apache-2.0.

Catalogue identity notes

Qwen/Qwen2-VL-2B-Instruct @895c3a4 (created 2024-08-28, licence apache-2.0). Smallest Qwen2-VL instruct model (vision encoder 675M + 1.5B LLM per report Table 1). Lab rows from the Qwen2-VL report (rule b) pooled with the catalogued Qwen2-VL 7B/72B and Qwen2.5-VL runs of first-party-qwen-scores; independent Vectara HHEM row. No ollama library tag (ollama ships qwen2.5vl only).

Model information sources

Release date ↗ · Weights licence ↗

HF safetensors metadata 2,208,985,600 parameters (Qwen/Qwen2-VL-2B-Instruct @895c3a4).

37.3±1.2#573 of 719 LLMsfrom 16 benchmark results across 2 sources

$ how it compares

The LLMs ranked either side of Qwen2-VL 2B Instruct, out of 828 with a score, then more from Alibaba.

#	model	score	where	RAM	released
664	Tulu 2 DPO 13B Allen Institute for AI	37±1	local	~9.8GB	2023-11
665	Yi-1.5 6B (base) 01.AI	37±3	local	~5.1GB	2024-05
666	Qwen2-VL 2B Instruct Alibaba	37±1	local	~2.8GB	2024-08
667	DeepSeek R1 Distill Qwen 1.5B DeepSeek	37±2	local	~2.6GB	2025-01
668	Gemma 2 2B IT Google	37±1	local	~3.1GB	2024-07
	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

math▁▃▃▅▆▅▄█▄▁▁▁ #387 of 398 * · 19% of its score · 2 tests

Trails Qwen2 0.5B Instruct, Granite 3.1 1B-A400M Instruct, Gemma 2B (base) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
MathVista (testmini) lab	43.0% ✎	▃··▃▅·▃█▅▅▅█ beats 0/16	medium · avg 69%	▇▇▇▇▇▇ 16%	at its level
MATH-Vision lab	12.4% ✎	▃▁·▁··▄▁▃▇▇█ beats 0/24	hard · avg 27%	▇▇ 4.0%	small run

vision & documents▃▁▃▂▁▃▄▂▆▇█▄ #145 of 147 · 85% of its score · 12 tests

Beats Gemma 3 4B, Phi-3-vision-128k-instruct here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
MMMU (val) lab	41.1% ✎	▃···▆▂▂▂▄▄▄█ beats 0/16	medium · avg 61%	▇▇▇▇▇▇▇ 16%	at its level
MathVista (testmini) lab	43.0% ✎	▃··▃▅·▃█▅▅▅█ beats 0/16	medium · avg 69%	▇▇▇▇▇▇ 16%	at its level
MMStar lab	48.0% ✎	▃··▄▂▂·▄█▂▄▆ beats 0/16	medium · avg 65%	▇▇▇▇▇ 12%	at its level
MMMU Pro lab	37.6% ✎	▃·█·▅▃▃▃▅▅▃█ beats 2/16	medium · avg 46%	▇▇▇ 7.0%	typical
DocVQA lab	90.1% ✎	▃··▅▃▅▃█▃█▅▅ beats 0/17	very easy · avg 94%	▇▇▇ 6.3%	blunt
ChartQA lab	73.5% ✎	▃··▂·▂▃·▅▂▆█ beats 0/17	easy · avg 85%	▇▇ 5.9%	blunt
InfographicVQA lab	65.5% ✎	▃·▃▅█▃▃▃▃▃▃▃ beats 3/13	easy · avg 80%	▇▇ 5.5%	at its level · small run
HallusionBench lab	41.7% ✎	▃·▃·▅▃▅▅▃▅█▅ beats 0/16	medium · avg 55%	▇▇ 5.2%	blunt · at its level

4 more tests

test	result	vs the others in its run	difficulty	counts	why
RealWorldQA lab	62.9% ✎	▅▃·▃▅▅·█·▃▅█ beats 0/16	easy · avg 70%	▇▇ 4.4%	blunt · at its level
MATH-Vision lab	12.4% ✎	▃▁·▁··▄▁▃▇▇█ beats 0/24	hard · avg 27%	▇▇ 4.0%	small run
OCRBench lab	80.9% ✎	▅█·▃▃█▃▅▅▃·▃ beats 3/16	easy · avg 84%	▇ 1.8%	blunt
TextVQA lab	79.7% ✎	█······▄·▄▄▄ beats 1/5	easy · avg 80%	▇ 1.0%	blunt · at its level

other tests4 not in a task · 15% of its score

test	result	vs the others in its run	difficulty	counts	why
Video-MME (without subtitles) lab	55.6% ✎	▃··█··▅█▃▃▅▃ beats 0/13	medium · avg 66%	▇▇▇ 6.9%	blunt · at its level
Video-MME (with subtitles) lab	60.4% ✎	▁▁▃▁▁·▁▃▁▁▁█ beats 2/16	medium · avg 73%	▇ 3.6%	at its level · small run
MVBench lab	63.2% ✎	▂▃▂·▂█··▆▄·▂ beats 1/13	medium · avg 67%	▇ 2.9%	blunt · at its level
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)	91.7% ✎	▁▁▂▃▄▄█▆▃▂▂▁ beats 33/140	very easy · avg 94%	▇ 1.3%	blunt · noisy

$ where to run it

llama.cpp · recommended

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 Qwen2-VL 2B Instruct · needs ~1.7 GB · from huggingface.co/ggml-org
chat in the terminal
llama cli -hf ggml-org/Qwen2-VL-2B-Instruct-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf ggml-org/Qwen2-VL-2B-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 Qwen2-VL 2B Instruct · needs ~1.7 GB · from huggingface.co/ggml-org
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/ggml-org/Qwen2-VL-2B-Instruct-GGUF@Q4_K_M -y
Then run lms ls to see its name, and lms chat followed by that name

Ollama

Install Ollama
download it (macOS / Windows) ↗
Linux
curl -fsSL https://ollama.com/install.sh | sh
Run Qwen2-VL 2B Instruct · needs ~1.7 GB · from huggingface.co/ggml-org
chat in the terminal
ollama run hf.co/ggml-org/Qwen2-VL-2B-Instruct-GGUF:Q4_K_M

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 Qwen2-VL 2B Instruct · needs ~1.2 GB, Mac only · from huggingface.co/mlx-community
chat in the terminal
mlx_lm.chat --model mlx-community/Qwen2-VL-2B-Instruct-4bit
Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model mlx-community/Qwen2-VL-2B-Instruct-4bit

Official weights and documentation: huggingface.co/Qwen ↗

$ how the score was worked out · Qwen (Alibaba) vendor-reporte… 98%, Vectara hallucination leaderb… 1%

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	15	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 98%	Factual vendor-reported score (not licensed content)
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)	1	▇ 1%	Apache-2.0
prior · pull toward the middle		▇ 1%
weighted average		= 37.3

$ 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