Qwen2.5-VL-7B-Instruct
Runs on your machine
Choose a machine
We then show the largest format that fits, from real file sizes where we have them.
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: 0.44 GB, worked out from the model's config (Qwen/Qwen2.5-VL-7B-Instruct), at 16-bit precision. Choose a machine above to see a verdict for each format. Method
weights context cache runtime
GGUF for llama.cpp, Ollama, LM Studio and others
unsloth/Qwen2.5-VL-7B-Instruct-GGUF (converted by unsloth), 295k downloads. Sizes are the real files on the Hub, all parts added up. For image input you also need the projector file mmproj-BF16.gguf (1.3 GB), not counted below. Showing the common levels.
| Format | Size | Memory needed | Memory | Get it |
|---|---|---|---|---|
| Q4_K_M | 4.4 GBfile size | 5.3 GB4.4 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q4_K_MOllama ollama run hf.co/unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q4_K_MDownload hf download unsloth/Qwen2.5-VL-7B-Instruct-GGUF --include "Qwen2.5-VL-7B-Instruct-Q4_K_M.gguf" --local-dir Qwen2.5-VL-7B-Instruct-GGUF | |
| Q5_K_M | 5.1 GBfile size | 6.0 GB5.1 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q5_K_MOllama ollama run hf.co/unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q5_K_MDownload hf download unsloth/Qwen2.5-VL-7B-Instruct-GGUF --include "Qwen2.5-VL-7B-Instruct-Q5_K_M.gguf" --local-dir Qwen2.5-VL-7B-Instruct-GGUF | |
| Q6_K | 5.8 GBfile size | 6.8 GB5.8 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q6_KOllama ollama run hf.co/unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q6_KDownload hf download unsloth/Qwen2.5-VL-7B-Instruct-GGUF --include "Qwen2.5-VL-7B-Instruct-Q6_K.gguf" --local-dir Qwen2.5-VL-7B-Instruct-GGUF | |
| Q8_0 | 7.5 GBfile size | 8.5 GB7.5 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q8_0Ollama ollama run hf.co/unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q8_0Download hf download unsloth/Qwen2.5-VL-7B-Instruct-GGUF --include "Qwen2.5-VL-7B-Instruct-Q8_0.gguf" --local-dir Qwen2.5-VL-7B-Instruct-GGUF | |
| BF16 | 14.2 GBfile size | 15.1 GB14.2 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen2.5-VL-7B-Instruct-GGUF:BF16Ollama ollama run hf.co/unsloth/Qwen2.5-VL-7B-Instruct-GGUF:BF16Download hf download unsloth/Qwen2.5-VL-7B-Instruct-GGUF --include "Qwen2.5-VL-7B-Instruct-BF16.gguf" --local-dir Qwen2.5-VL-7B-Instruct-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Memory | Repo |
|---|---|---|---|---|
| MLX 4bitmoot20/Qwen2.5-VL-7B-Instruct-MLX-4bits | 4.5 GBfile size | 5.5 GB4.5 + 0.44 + 0.50 | Download page | |
| MLX 8bitmoot20/Qwen2.5-VL-7B-Instruct-MLX-8bits | 8.3 GBfile size | 9.3 GB8.3 + 0.44 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Memory | Repo |
|---|---|---|---|---|
| BF16original weights | 15.4 GBfile size | 16.4 GB15.4 + 0.44 + 0.50 |
hf download Qwen/Qwen2.5-VL-7B-Instruct --local-dir Qwen2.5-VL-7B-InstructOther conversions (15; sizes not read)
- Qwen/Qwen2.5-VL-7B-Instruct-AWQ AWQ, by the maker, 712k downloads
- nvidia/Qwen2.5-VL-7B-Instruct-NVFP4 NVFP4, 784k downloads
- RedHatAI/Qwen2.5-VL-7B-Instruct-FP8-Dynamic FP8, 118k downloads
- RedHatAI/Qwen2.5-VL-7B-Instruct-quantized.w8a8 Other, w8a8, 19k downloads
- unsloth/Qwen2.5-VL-7B-Instruct-bnb-4bit bitsandbytes, 4bit, 8.8k downloads
- unsloth/Qwen2.5-VL-7B-Instruct-unsloth-bnb-4bit bitsandbytes, 4bit, 6.6k downloads
- nvidia/Qwen2.5-VL-7B-Instruct-FP8 FP8, 1.2k downloads
- hfl/Qwen2.5-VL-7B-Instruct-GPTQ-Int4 GPTQ, int4, 742 downloads
- OpenVINO/Qwen2.5-VL-7B-Instruct-int4-ov Other, int4, 661 downloads
- RedHatAI/Qwen2.5-VL-7B-Instruct-quantized.w4a16 Other, w4a16, 618 downloads
- OpenVINO/Qwen2.5-VL-7B-Instruct-int8-ov Other, int8, 418 downloads
- ig1/Qwen2.5-VL-7B-Instruct-FP8-Dynamic FP8, 374 downloads
- ig1/Qwen2.5-VL-7B-Instruct-NVFP4 NVFP4, 91 downloads
- wanzhenchn/Qwen2.5-VL-7B-Instruct-gptqmodel-int8 GPTQ, int8, 49 downloads
- leon-se/Qwen2.5-VL-7B-Instruct-FP8-Dynamic FP8, 40 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
- huihui-ai/Qwen2.5-VL-7B-Instruct-abliterated8.3B309kUnreviewed
- allenai/olmOCR-2-7B-10258.3B81kUnreviewed
- mlx-community/Qwen2.5-VL-7B-Instruct-4bit18kUnreviewed
- nvidia/Cosmos-Reason1-7B8.3B16kUnreviewed
- xlangai/OpenCUA-7B8.3B8.0kUnreviewed
- Vchitect/ShotVL-7B8.3B6.2kUnreviewed
Show 13 moreShow fewer
- unsloth/Qwen2.5-VL-7B-Instruct8.3B4.1kUnreviewed
- zhiyuanyou/Qwen2.5-VL-7B-GRPO-Composition-Score-Class8.3B2.7kUnreviewed
- microsoft/X-Reasoner-7B8.3B2.3kUnreviewed
- openbmb/EVisRAG-7B8.3B2.3kUnreviewed
- baichuan-inc/BaichuanMed-OCR-7B8.3B2.2kUnreviewed
- internlm/Spatial-SSRL-7B8.3B2.0kUnreviewed
- inclusionAI/ZwZ-7B8.3B1.9kUnreviewed
- OpenGVLab/ScaleCUA-7B8.3B1.8kUnreviewed
- inclusionAI/ViLaSR8.3B1.8kUnreviewed
- allenai/olmOCR-7B-07258.3B775Unreviewed
- allenai/olmOCR-7B-08258.3B692Unreviewed
- microsoft/GUI-Actor-7B-Qwen2.5-VL8.4B435Unreviewed
- sensenova/SenseNova-SI-1.1-Qwen2.5-VL-7B8.3B102Unreviewed
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.