- Size
- 4.4B parameterscounted from the safetensors header on the Hub
- Licence
- unknown, Apache License 2.0Our reading, check the licence: You can use, change and redistribute the model, including commercially. Keep the licence and any notices, and say which files you changed; you do not have to publish your changes. It includes a patent licence that ends for anyone who sues over patents in the work.
- On the Hub since
- 11 Oct 2025the date the repo was created
- Context length
- 256k tokensfrom the model's config
- Kind
- Base model (no base model declared on the Hub)Qwen3VLForConditionalGeneration
- On Hugging Face
- Qwen/Qwen3-VL-4B-Instruct3.0M downloads in the last 30 days, 541 likes
Scores
No independent scores for this model in Epoch AI's data. A score is only shown here when Epoch's name for a model matches this repo with certainty; an uncertain match is left out rather than guessed.
Claimed by the maker Claimed
The model card has no claimed scores in a machine-readable form (the card's model-index), so none are shown. The maker may give scores in the card text: read the card.
Other views, which we link to and do not copy: look this model up on Artificial Analysis and the Arena leaderboard.
Run it yourself
Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 1.1 GB, worked out from the model's config (Qwen/Qwen3-VL-4B-Instruct), at 16-bit precision. Choose a machine above to see a verdict for each format. Method
GGUF for llama.cpp, Ollama, LM Studio and others
Qwen/Qwen3-VL-4B-Instruct-GGUF (the maker's own conversion), 630k downloads. Sizes are the real files on the Hub, all parts added up. For image input you also need the projector file mmproj-Qwen3VL-4B-Instruct-F16.gguf (0.78 GB), not counted below.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| Q4_K_M | 2.3 GBfile size | 4.0 GB2.3 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-VL-4B-Instruct-GGUF:Q4_K_MOllama ollama run hf.co/Qwen/Qwen3-VL-4B-Instruct-GGUF:Q4_K_MDownload hf download Qwen/Qwen3-VL-4B-Instruct-GGUF --include "Qwen3VL-4B-Instruct-Q4_K_M.gguf" --local-dir Qwen3-VL-4B-Instruct-GGUF |
| Q8_0 | 4.0 GBfile size | 5.6 GB4.0 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-VL-4B-Instruct-GGUF:Q8_0Ollama ollama run hf.co/Qwen/Qwen3-VL-4B-Instruct-GGUF:Q8_0Download hf download Qwen/Qwen3-VL-4B-Instruct-GGUF --include "Qwen3VL-4B-Instruct-Q8_0.gguf" --local-dir Qwen3-VL-4B-Instruct-GGUF |
| F16 | 7.5 GBfile size | 9.1 GB7.5 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-VL-4B-Instruct-GGUF:F16Ollama ollama run hf.co/Qwen/Qwen3-VL-4B-Instruct-GGUF:F16Download hf download Qwen/Qwen3-VL-4B-Instruct-GGUF --include "Qwen3VL-4B-Instruct-F16.gguf" --local-dir Qwen3-VL-4B-Instruct-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLX 4bitlmstudio-community/Qwen3-VL-4B-Instruct-MLX-4bit | 2.9 GBfile size | 4.5 GB2.9 + 1.1 + 0.50 | Download page |
| MLX 8bitlmstudio-community/Qwen3-VL-4B-Instruct-MLX-8bit | 4.8 GBfile size | 6.4 GB4.8 + 1.1 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| BF16original weights | 8.3 GBfile size | 9.9 GB8.3 + 1.1 + 0.50 |
hf download Qwen/Qwen3-VL-4B-Instruct --local-dir Qwen3-VL-4B-InstructOther conversions (20; sizes not read)
- Qwen/Qwen3-VL-4B-Instruct-FP8 FP8, by the maker, 167k downloads
- mobilint/Qwen3-VL-4B-Instruct Other, 27k downloads
- unsloth/Qwen3-VL-4B-Instruct-unsloth-bnb-4bit bitsandbytes, 4bit, 14k downloads
- cyankiwi/Qwen3-VL-4B-Instruct-AWQ-4bit AWQ, 4bit, 5.0k downloads
- unsloth/Qwen3-VL-4B-Instruct-bnb-4bit bitsandbytes, 4bit, 2.0k downloads
- JEILDLWLRMA/Qwen3-VL-4B-Instruct-NVFP4 NVFP4, 956 downloads
- OpenVINO/Qwen3-VL-4B-Instruct-int4-ov Other, int4, 413 downloads
- cyankiwi/Qwen3-VL-4B-Instruct-AWQ-8bit AWQ, 8bit, 241 downloads
- OpenVINO/Qwen3-VL-4B-Instruct-int4-gq-ov Other, int4, 112 downloads
- taobao-mnn/Qwen3-VL-4B-Instruct-MNN Other, 109 downloads
- MLliu6/Qwen3-VL-4B-Instruct-GPTQ-W4A16 AWQ, w4a16, 91 downloads
- robotflowlabs/qwen3-vl-4b-instruct-int4 bitsandbytes, int4, 88 downloads
- tgpm/Qwen3-VL-4B-Instruct-FP8 FP8, 74 downloads
- JEILDLWLRMA/Qwen3-VL-4B-Instruct-FP8-Dynamic FP8, 57 downloads
- simaai/Qwen3-VL-4B-Instruct-Autoround-Safetensors GPTQ, 32 downloads
- MLliu6/Qwen3-VL-4B-Instruct-SmoothQuant-W8A8-FP8 AWQ, w8a8, 28 downloads
- ArtusDev/Qwen_Qwen3-VL-4B-Instruct-EXL3 EXL3, 27 downloads
- AutomatosX/AX-Qwen3-VL-4B-Instruct-CUDA-AXQ-NVFP4-W4A4 NVFP4, 26 downloads
- sdr001/Qwen3-VL-4B-Instruct-FP8 FP8, 22 downloads
- pramjana/Qwen3-VL-4B-Instruct-4bit-GPTQ GPTQ, 4bit, 20 downloads
The Hub tags 100 or more quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.
Run it by API
No provider on Hugging Face Inference Providers serves this model (as of 19 Nov 2026, 13:58 UTC). Other hosts may; we only list what the router reports.
Not built yet: one key for every open model, with BlockRun, and our own tested-quality score per provider. Until then, precision and measured speed are the best guide to a provider here.
Fine-tunes of this model
Repos that declare this model as their base, most downloaded first. Listing is not a recommendation; unreviewed ones have not been checked.
- unsloth/Qwen3-VL-4B-Instruct4.4B16kUnreviewed
- huihui-ai/Huihui-Qwen3-VL-4B-Instruct-abliterated4.4B6.8kUnreviewed
- stepfun-ai/GELab-Zero-4B-preview4.4B2.2kUnreviewed
- internlm/Spatial-SSRL-Qwen3VL-4B4.8B1.7kUnreviewed
- inclusionAI/ZwZ-4B4.8B1.6kUnreviewed
Planned for this page Not built yet
- Try it in private chat
- A link to Solana AI's confidential chat for the models it serves. This model is not listed as served there.
- Fine-tune it
- Fine-tuning a few base models from this page and publishing the result here.
- Creators and earnings
- Pages for people who publish fine-tunes and adapters of this model, and what they earn.
- Fund and bounties
- Bounties for new open models, paid from a public fund. How the fund will work.
- Contribute compute
- How to give compute to open model work, with Akash.
Model data from the Hugging Face Hub 19 Nov 2026, 13:43 UTC. The repo. Found a mistake on this page? Tell us.