Alternatives to Qwen2.5-VL-72B-Instruct

Qwen2.5-VL-72B-Instruct alternatives

Qwen2.5-VL-72B-Instruct's smallest published quantization is 22.11 GiB. The models below do the same job with a different trade — less memory, comparable size, or a more permissive licence.

From the file· sizes from summed file bytes

Meaningfully smaller

under 70% of its smallest quantization
ModelParamsSmallestQuantsLicence
Qwen3.6-35B-A3BMoE36.0B8.77 GiB61apache-2.0
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP27.8B10.12 GiB34apache-2.0
Qwen3.5-9B9.7B2.97 GiB34apache-2.0
gemma-4-26B-A4B-itMoE26.5B8.99 GiB46apache-2.0
gemma-4-12B-it12.0B3.92 GiB59apache-2.0
Qwen3.5-4B4.7B1.42 GiB57apache-2.0
Qwythos-9B-Claude-Mythos-5-1M9.4B5.38 GiB16apache-2.0
gemma-4-31B-it31.3B7.95 GiB55apache-2.0

Comparable in size

within ±40%, so a like-for-like swap
ModelParamsSmallestQuantsLicence
Qwen3.5-122B-A10BMoE125B26.92 GiB58apache-2.0
gemma-4-31B-it-qat-q4_0-unquantized32.7B16.44 GiB1apache-2.0
Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-PreservedMoE35.1B16.08 GiB10apache-2.0
Qwen3.6-35B-A3B-uncensored-hereticMoE35.1B15.71 GiB10apache-2.0
Llama-4-Scout-17B-16E-InstructMoE109B24.51 GiB48other
UI-TARS-72B-DPO73.4B22.11 GiB26apache-2.0

More permissively licensed

licences that allow commercial use
ModelParamsSmallestQuantsLicence
Qwen3.6-35B-A3BMoE36.0B8.77 GiB61apache-2.0
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP27.8B10.12 GiB34apache-2.0
Qwen3.5-9B9.7B2.97 GiB34apache-2.0
gemma-4-26B-A4B-itMoE26.5B8.99 GiB46apache-2.0
gemma-4-12B-it12.0B3.92 GiB59apache-2.0
Qwen3.5-4B4.7B1.42 GiB57apache-2.0
Qwythos-9B-Claude-Mythos-5-1M9.4B5.38 GiB16apache-2.0
gemma-4-31B-it31.3B7.95 GiB55apache-2.0