Molmo 7B-D

Molmo 7B-D needs an estimated 6.1 GB of VRAM at Q4, 10.3 GB at Q8 and 18.9 GB at FP16. All figures are estimates based on 8B parameters, weights plus about 15% and 0.5 GB for a short context.

Yes. A 24 GB GPU fits FP16 at an estimated 18.9 GB, a 16 GB or 12 GB GPU fits Q8 at an estimated 10.3 GB, and an 8 GB GPU fits Q4 at an estimated 6.1 GB.

Yes. With Q4 needing an estimated 6.1 GB and macOS letting the GPU use about 75% of unified memory, the smallest common Mac memory size that fits Q4 is an estimated 16 GB.

Yes. Molmo 7B-D has open weights, so you can self-host it for full control of your data.

Run Molmo 7B-D locally

Molmo 7B-D accepts text and image inputs. It does not accept audio, video or PDFs.

Molmo 7B-D ranks 158 on Arena Vision, Arena's own rank, with a score of 994.

Molmo 7B-D benchmarks

Matched on model kind and release date, with no comparable price, the closest alternatives are Llama-3.2-11B-Vision-Instruct and Llama-3.2-1B by Meta, and Qwen2.5 7B Instruct by Alibaba.

Llama-3.2-11B-Vision-InstructMolmo 7B-D vs Llama-3.2-11B-Vision-InstructLlama-3.2-1BMolmo 7B-D vs Llama-3.2-1BQwen2.5 7B InstructMolmo 7B-D vs Qwen2.5 7B Instruct

Molmo 7B-D has open weights. It is released under the Apache 2.0 license.

Yes. Molmo 7B-D uses the Apache 2.0 license, a permissive license that allows commercial use.

The Molmo 7B-D weights are available in the allenai/Molmo-7B-D-0924 repository on Hugging Face.

Run Molmo 7B-D locally

Molmo 7B-D was made by Ai2 and released on September 25, 2024.

More from Ai2

Yes. Molmo 7B-D is an open-weights model, so it is free to use.