whatmodel.sh / Qwen3-VL 2B Instruct

Qwen3-VL 2B Instruct

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Qwen3-VL 2B Instruct is an open-weights language model from Alibaba, released October 2025. Its published size is 2.13B parameters and a 256k-token context window. Licence: apache-2.0.

Catalogue identity notes

Maker repo Qwen/Qwen3-VL-2B-Instruct @8964489 (licence tag apache-2.0, created 2025-10-19); the separate Instruct checkpoint, not Qwen3-VL-2B-Thinking. Scored from the Qwen3-VL Technical Report (arXiv 2511.21631v2, Tables 3/4 and 7-10: column 'Qwen3-VL 2B 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 ↗

2,127,532,032 parameters in the Hugging Face safetensors metadata of Qwen/Qwen3-VL-2B-Instruct @8964489 (vision encoder included); dense, so active = total.

43.9±0.7#464 of 719 LLMsfrom 41 benchmark results across 5 sources

$ how it compares

The LLMs ranked either side of Qwen3-VL 2B Instruct, out of 828 with a score, then more from Alibaba.

#modelscorewhereRAMreleased
551Falcon3 7B Instruct TII44±3local~5.9GB2024-11
552Xwin-LM 70B v0.1 Xwin-LM44±1local~28GB2023-09
553Qwen3-VL 2B Instruct Alibaba44±1local~2.7GB2025-10
554Dolphin 2.6 Mixtral 8x7B Cognitive Computations44±1local~29GB2023-12
555Granite 8B Code Instruct (4K) IBM44±1local~6.6GB2024-04
more from Alibaba
18Qwen3.8 Max (2.4T-A95B) Alibaba78±1api $3.00/M +local~380GB2026-08
34Qwen3.8 Max (0902) Alibaba76±1api $3.00/M2026-09
37Qwen3.8-Flash-Next Alibaba75±1local~59GB2026-08

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #235 of 300 * · 1.8% of its score · 1 test

Beats o3-mini, Ministral 3 14B Reasoning 2512, DeepSeek R1 0528 Qwen3 8B here, though they rank higher overall. Trails Granite 4.0 1B, SmolLM3 3B, Phi-3.5-mini-instruct here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
BFCL v3 lab55.4% ✎█▄··▄█·█▄█▄█ beats 0/13medium · avg 59%▇ 1.8%blunt · at its level
coding▁▁▁▃▆██▇▆▄▃▁ #414 of 458 * · 2.6% of its score · 1 test

Beats DeepSeek R1 Distill Qwen 7B, Qwen3.5 2B, DeepSeek R1 Distill Llama 8B here, though they rank higher overall. Trails Llama 3.2 1B Instruct, SmolLM2 1.7B Instruct, Granite 3B Code Base (2K) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
LiveCodeBench v6 (window unspecified) lab20.3% ✎▃▅█▅▅▃▅···▃█ beats 0/16medium · avg 35%▇ 2.6%typical
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #279 of 393 * · 4.2% of its score · 1 test

Beats Ministral 3 14B Reasoning 2512, Gemma 4 E4B, Ministral 3 8B here, though they rank higher overall. Trails Flan-T5 XXL, LFM2 700M, EXAONE 3.5 2.4B Instruct here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
IFEval (aggregation unspecified) lab +168.2% ✎▃··▃···██▅▃▃ beats 0/11easy · avg 77%▇▇ 4.2%at its level · several runs
math▁▂▃▅▆▅▄█▄▁▁▁ #290 of 398 * · 9.1% of its score · 2 tests

Beats GPT-4 (0613), Athene-70B, Gemma 2 27B IT here, though they rank higher overall. Trails DeepSeek R1 Distill Qwen 1.5B, Phi-2 (2.7B), StarCoder2 15B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
MathVista (testmini) lab61.3% ✎▃··▃▅·▃█▅▅▅█ beats 2/16medium · avg 69%▇▇ 5.6%at its level
AIME 2025 lab22.2% ✎▃·▃▃█··▃▅▃·█ beats 0/12medium · avg 45%▇ 3.4%sharp · at its level
vision & documents▁▁▃▂▁▃▄▂▆▇█▄ #131 of 147 · 53% of its score · 19 tests

Beats Gemma 4 E2B, LLaVA-OneVision Qwen2 72B (ov-sft), Gemma 3 4B here, though they rank higher overall.

testresultvs the others in its rundifficultycountswhy
CharXiv descriptive questions lab +162.3% ✎▄··▂▂▂·▂▄▄▂█ beats 1/14easy · avg 74%▇▇▇▇ 11%at its level · several runs
MMMU (val) lab53.4% ✎▂···▆▂▂▂▄▄▄█ beats 2/16medium · avg 61%▇▇▇ 6.6%at its level
CharXiv (reasoning questions) lab26.8% ✎▄▂▄▆▄▄▆··▄█▂ beats 0/21medium · avg 39%▇▇▇ 6.3%at its level
MMStar lab +158.3% ✎▂··▄▂▂·▄█▂▄▆ beats 2/16medium · avg 65%▇▇ 6.2%at its level · several runs
MathVista (testmini) lab61.3% ✎▃··▃▅·▃█▅▅▅█ beats 2/16medium · avg 69%▇▇ 5.6%at its level
MMMU Pro lab36.5% ✎▃·█·▅▃▃▃▅▅▃█ beats 1/16medium · avg 46%▇ 3.1%at its level
HallusionBench lab +151.4% ✎▃·▃·▅▃▅▅▃▅█▅ beats 3/16medium · avg 55%▇ 2.6%blunt · at its level · several runs
ChartQA lab +179.1% ✎▂··▃·▂▃·▅▂▆█ beats 1/17easy · avg 85%▇ 1.9%blunt · several runs
11 more tests
testresultvs the others in its rundifficultycountswhy
RealWorldQA lab +163.9% ✎▅▃·▃▅▅·█·▃▅█ beats 1/16easy · avg 70%▇ 1.8%blunt · at its level · several runs
DocVQA lab93.3% ✎▃··▅▃▅▃█▃█▅▅ beats 2/17very easy · avg 94%▇ 1.7%blunt · maxed out
ChartArena chart parsing (English, mAP_high average)27.0% ✎▁·▁·▃▃▁▇▃█·▄ beats 5/22hard · avg 27%▇ 1.6%blunt
ScreenSpot-Pro lab48.5% ✎▂····▂▄··▂█▆ beats 2/11medium · avg 46%▇ 1.0%blunt · at its level
CC-OCR lab72.8% ✎▃··▃▅·▅▅·█▅▃ beats 1/13easy · avg 74%▇ 1.0%blunt · at its level
MMBench-EN dev v1.174.4% ✎█·····▃··▃·▃ beats 4/5easy · avg 79%▇ 1.0%blunt · at its level
BEAR-Bench (bilingual RU/EN enterprise and academic document images, overall accuracy)14.3% ✎▃▃·▅·▃·█·█▅▅ beats 1/14hard · avg 24%▇ 0.9%typical
OCRBench lab +185.8% ✎▅█·▃▃█▃▅▅▃·▃ beats 7/16easy · avg 84%▇ 0.9%blunt · several runs
CountBench84.8% ✎▄··▄··█··▄·▄ beats 5/5very easy · avg 86%▇ 0.4%noisy
AI2D test71.3% ✎█···▄▄···▄·▄ beats 1/5easy · avg 82%▇ 0.2%blunt · at its level
PureDocBench document parsing (Avg3 over Clean/Digital/Real)62.1% ✎▂·▁·▃▂▅█·▁▂▂ beats 3/22medium · avg 62%▇ 0.1%blunt · at its level
other tests10 not in a task · 34% of its score
testresultvs the others in its rundifficultycountswhy
MMLU-Pro lab +149.0% ✎▄·▄·▄▄▄▄█▄██ beats 0/12medium · avg 63%▇▇▇▇ 9.0%at its level · several runs
MMLU-Redux lab66.5% ✎▃··▃·▃·█▃▅▃▅ beats 0/11easy · avg 79%▇▇▇ 7.0%at its level
GPQA (subset unspecified) lab42.0% ✎▂····▃·▂▂█▆█ beats 1/14hard · avg 49%▇▇ 4.6%sharp
SuperGPQA lab24.3% ✎▄··▄·▄▄▄█▄██ beats 0/11medium · avg 37%▇ 3.0%at its level
Video-MME (without subtitles) lab61.9% ✎▃··█··▅█▃▃▅▃ beats 1/13medium · avg 66%▇ 2.5%blunt · at its level
AndroidWorld lab36.4% ✎▃··▃▅▃▃▃▃▃·█ beats 0/11medium · avg 43%▇ 2.4%blunt · at its level
MMLU63.0% ✎▄·····▄▄··▄█ beats 3/5medium · avg 70%▇ 2.1%at its level
LVBench lab47.4% ✎▅··▃·▅▃▃·█·▃ beats 0/10hard · avg 51%▇ 2.0%at its level
2 more tests
testresultvs the others in its rundifficultycountswhy
MVBench lab61.7% ✎▃▂▂·▂█··▆▄·▂ beats 0/13medium · avg 67%▇ 1.0%blunt · at its level
Multi-IF (aggregation unspecified)52.3% ✎▄·▄··█▄····▄ beats 4/5medium · avg 58%▇ 0.9%blunt · at its level

$ where to run it

llama.cpp · recommended

version

  1. 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 ↗

  2. Run Qwen3-VL 2B Instruct · needs ~1.1 GB · from huggingface.co/Qwen (official)
    chat in the terminal
    llama cli -hf Qwen/Qwen3-VL-2B-Instruct-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf Qwen/Qwen3-VL-2B-Instruct-GGUF:Q4_K_M

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run Qwen3-VL 2B Instruct · needs ~1.1 GB · from huggingface.co/Qwen (official)

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/Qwen/Qwen3-VL-2B-Instruct-GGUF@Q4_K_M -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run Qwen3-VL 2B Instruct · needs ~1.9 GB · from ollama.com
    chat in the terminal
    ollama run qwen3-vl:2b-instruct

MLX (Mac)

version

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run Qwen3-VL 2B Instruct · needs ~1.8 GB, Mac only · from huggingface.co/mlx-community
    chat in the terminal
    mlx_lm.chat --model mlx-community/Qwen3-VL-2B-Instruct-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model mlx-community/Qwen3-VL-2B-Instruct-4bit

Official weights and documentation: huggingface.co/Qwen ↗

$ how the score was worked out · Qwen (Alibaba) vendor-reporte… 84%, CohereLabs/North-Micro-Vision… 13%, ChartArena: Benchmarking Char… 2% and 2 more

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

sourceresultsshare of the scorelicence
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-reported25▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 84%Factual vendor-reported score (not licensed content)
CohereLabs/North-Micro-Vision-Instruct model card (benchmark table)13▇▇▇▇ 13%Apache-2.0
ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)1▇ 2%CC BY 4.0
BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)1▇ 1%CC BY 4.0
How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)1▇ 0%CC BY 4.0
prior · pull toward the middle▇ 0%
weighted average= 43.9

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit 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 · sources and licences · json