[whatmodel.sh](../../README.md) / Qwen3-VL 2B Thinking # Qwen3-VL 2B Thinking Alibaba · released 2025-10 · 2.13B params · apache-2.0 **score 49.2** ±0.7 · #388 of 732 · from 26 benchmark results across 2 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at _{● marks Qwen3-VL 2B Thinking among every model on that task or test, worst to best.}

task                vs the others          rank  counts  tests
agent tasks         ▁▁▃▆●▄▅▅▄▄▂▁   #240 of 313*    2.1%      1
coding              ▁▁▁▃●▇█▇▅▄▃▁   #382 of 469*    3.5%      1
chat & writing      ▁▁▁▃▆▄●█▆▇▇▃   #260 of 402*    3.4%      1
math                ▁▁▃▅█●▄█▄▂▁▁   #225 of 431*     11%      2
vision & documents  ▁▁▂▂●▃▄▂▆▇█▄    #126 of 153     51%     14

agent tasks · #240 of 313 · 2.1% of its score

Beats [o3-mini](o3-mini.md), [Nova 2 Lite](nova-2-lite.md), [Llama 4 Maverick (402B-A17B)](llama-4-maverick.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Phi-3.5-mini-instruct](phi-3-5-mini-instruct.md), [Jamba 1.5 Mini](jamba-1-5-mini.md) here, though they rank lower overall.

test             result  vs the others in its run  difficulty        counts  why
BFCL v3 (maker)   57.2%  █●··▄█·█▄█▄█ beats 2/13   medium · avg 59%    2.1%  blunt · at its level

coding · #382 of 469 · 3.5% of its score

Beats [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md) here, though they rank higher overall. Trails [Qwen2.5 1.5B Instruct](qwen2-5-1-5b.md), [WizardLM 13B V1.2](wizardlm-13b-v1-2.md), [Granite 3B Code Instruct (2K)](granite-3b-code-instruct-2k.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty      counts  why
LiveCodeBench v6 (window unspecif…   29.3%  ▃●█▅▅▃▅···▃█ beats 1/16   hard · avg 35%    3.5%  at its level

chat & writing · #260 of 402 · 3.4% 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), [Granite 4.0 1B](granite-4-0-1b.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty      counts  why
IFEval (aggregation unspecified) …   75.1%  ▃··●···██▅▃▃ beats 1/11   easy · avg 77%    3.4%  typical

math · #225 of 431 · 11% 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
MathVista (testmini) (maker)   73.6%  ▃··▃▅·▃●▅▅▅█ beats 6/16   medium · avg 69%    5.8%  at its level
AIME 2025 (maker)              39.0%  ▃·▃●█··▃▅▃·█ beats 2/12   medium · avg 44%    4.7%  sharp · at its level

vision & documents · #126 of 153 · 51% of its score

Beats [DeepSeek V4 Flash Vision (Exp)](deepseek-v4-flash-vision-exp.md), [Gemma 3 12B](gemma-3-12b.md), [Llama 3.2 90B Vision Instruct](llama-3-2-90b-vision.md) here, though they rank higher overall. Trails [Qwen3.5 2B](qwen3-5-2b.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty           counts  why
CharXiv (reasoning questions) (ma…   37.1%  ▄●▄▆▄▄▆··▄█▂ beats 2/21   medium · avg 38%       8.6%  at its level
CharXiv descriptive questions (ma…   70.1%  ▄··●▂▂·▂▄▄▂█ beats 2/14   easy · avg 74%         8.3%  at its level
MMMU (val) (maker)                   61.4%  ▂···▆▂●▂▄▄▄█ beats 5/16   medium · avg 61%       7.6%  at its level
MathVista (testmini) (maker)         73.6%  ▃··▃▅·▃●▅▅▅█ beats 6/16   medium · avg 69%       5.8%  at its level
MMStar (maker)                       68.1%  ▂··▄▂▂·●█▂▄▆ beats 5/16   medium · avg 65%       5.1%  at its level
MMMU Pro (maker)                     42.5%  ▃·█·●▃▃▃▅▅▃█ beats 4/16   medium · avg 46%       3.9%  at its level
ChartArena chart parsing (English…   26.5%  ▁·▁·▃●▁▇▃█·▄ beats 4/22   hard · avg 27%         2.1%  blunt · at its level
HallusionBench (maker)               54.9%  ▃·▃·▅▃●▅▃▅█▅ beats 5/16   medium · avg 54%       2.1%  blunt · at its level
ChartQA (maker)                      86.6%  ▂··▂·▂▃·●▂▆█ beats 5/17   easy · avg 85%         1.8%  blunt
DocVQA (maker)                       92.9%  ▃··●▃▅▃█▃█▅▅ beats 1/17   very easy · avg 94%    1.7%  blunt · maxed out
RealWorldQA (maker)                  69.5%  ▅▃·▃●▅·█·▃▅█ beats 4/16   easy · avg 70%         1.6%  blunt · at its level
ScreenSpot-Pro (maker)               32.2%  ●····▂▄··▂█▆ beats 0/11   medium · avg 46%       1.2%  blunt · at its level
CC-OCR (maker)                       68.3%  ●··▃▅·▅▅·█▅▃ beats 0/13   easy · avg 74%         1.0%  blunt · at its level
OCRBench (maker)                     79.2%  ●█·▃▃█▃▅▅▃·▃ beats 0/16   easy · avg 84%         0.6%  blunt

other tests (8)

test                                result  vs the others in its run  difficulty        counts  why
MMLU-Pro (maker)                     62.3%  ▄·▄·●▄▄▄█▄██ beats 2/12   medium · avg 63%    7.9%  at its level
GPQA (subset unspecified) (maker)    49.5%  ▂····▂·●▂█▆█ beats 2/14   hard · avg 49%      7.0%  sharp · at its level
MMLU-Redux (maker)                   76.9%  ▃··●·▃·█▃▅▃▅ beats 1/11   easy · avg 79%      6.8%  at its level
SuperGPQA (maker)                    34.6%  ▄··●·▄▄▄█▄██ beats 1/11   medium · avg 37%    3.6%  at its level
AndroidWorld (maker)                 46.1%  ▃··▃●▃▃▃▃▃·█ beats 2/11   medium · avg 43%    2.9%  at its level
Video-MME (without subtitles) (ma…   62.1%  ▃··●··▅█▃▃▅▃ beats 2/13   medium · avg 66%    2.8%  blunt · at its level
LVBench (maker)                      47.6%  ●··▃·▅▃▃·█·▃ beats 1/10   hard · avg 51%      2.5%  at its level
MVBench (maker)                      64.5%  ▂▂●·▂█··▆▄·▂ beats 2/13   medium · avg 67%    1.1%  blunt · at its level

## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's 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 is added up by who ran the tests. [Method](../method.md)

source                                    results  share of the score
Qwen (submitted / maker-reported)              25   98% ▇▇▇▇▇▇▇▇▇▇▇▇
ChartArena: Benchmarking Chart Parsing …        1    2% ▇
prior: pull toward the middle                        0% ▇
weighted average                                   = 49.2

_{Licences: ChartArena: Benchmarking Chart Parsing across Languages ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/))} ## 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-VL 2B Thinking (Q4_K_M, needs ~1.1 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-VL-2B-Thinking-GGUF/tree/76646427920e08bd278e46af11c5db8532e6a657)) ```sh llama cli -hf Qwen/Qwen3-VL-2B-Thinking-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-VL-2B-Thinking-GGUF:Q4_K_M ``` Other versions: Q8_0 (~1.8 GB) `llama cli -hf Qwen/Qwen3-VL-2B-Thinking-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-VL 2B Thinking (Q4_K_M, needs ~1.1 GB, from [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-VL-2B-Thinking-GGUF/tree/76646427920e08bd278e46af11c5db8532e6a657)) ```sh lms get https://huggingface.co/Qwen/Qwen3-VL-2B-Thinking-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 (~1.8 GB) `lms get https://huggingface.co/Qwen/Qwen3-VL-2B-Thinking-GGUF@Q8_0 -y` · 4bit (~1.8 GB, Mac only) `lms get https://huggingface.co/mlx-community/Qwen3-VL-2B-Thinking-4bit -y` · 8bit (~2.6 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/Qwen3-VL-2B-Thinking-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-VL 2B Thinking (Q4_K_M, needs ~1.9 GB, from [ollama.com](https://ollama.com/library/qwen3-vl/tags)) ```sh ollama run qwen3-vl:2b-thinking ``` Other versions: Q8_0 (~2.6 GB) `ollama run qwen3-vl:2b-thinking-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-VL 2B Thinking (4bit, needs ~1.8 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/Qwen3-VL-2B-Thinking-4bit/tree/6c666f518757f9378f581ea0252822fe8b1c5131)) ```sh mlx_lm.chat --model mlx-community/Qwen3-VL-2B-Thinking-4bit ``` 3. Or start an OpenAI-compatible server at http://localhost:8080/v1 ```sh mlx_lm.server --model mlx-community/Qwen3-VL-2B-Thinking-4bit ``` Other versions: 8bit (~2.6 GB, Mac only) `mlx_lm.chat --model lmstudio-community/Qwen3-VL-2B-Thinking-MLX-8bit` Model repository and documentation: [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-VL-2B-Thinking) [interactive version](https://whatmodel.sh/model/qwen3-vl-2b-thinking) · [JSON](https://whatmodel.sh/api?format=json&all=1)