MolmoPoint 8B

MolmoPoint 8B has 8.7B parameters. Estimated VRAM is 6.6 GB at Q4, 11.1 GB at Q8 and 20.5 GB at FP16. All figures are estimates for weights plus about 15% and 0.5 GB for a short context.

Yes. A 24 GB GPU fits at FP16 (estimated 20.5 GB), a 16 GB or 12 GB GPU fits at Q8 (estimated 11.1 GB) and an 8 GB GPU fits at Q4 (estimated 6.6 GB).

Yes. At Q4 the estimated VRAM is 6.6 GB. Assuming macOS lets the GPU use about 75% of unified memory, the smallest common Mac memory size that fits is 16 GB.

Yes. MolmoPoint 8B has open weights, so you can self-host it for full control of your data.

Run MolmoPoint 8B locally

It can read images but not audio, video or PDFs. Its supported input modalities are text and image.

Matched on model kind and release date (no comparable price), the closest alternatives are MiniMax-M2.7-highspeed by MiniMax, Nemotron Cascade 2 30B A3B by NVIDIA and Nemotron 3 Content Safety by NVIDIA.

MiniMax-M2.7-highspeedMolmoPoint 8B vs MiniMax-M2.7-highspeedNemotron Cascade 2 30B A3BMolmoPoint 8B vs Nemotron Cascade 2 30B A3BNemotron 3 Content SafetyMolmoPoint 8B vs Nemotron 3 Content Safety

MolmoPoint 8B has open weights, not fully open source. The license is Apache 2.0.

Yes. The license is Apache 2.0, a permissive license that allows commercial use.

The weights can be downloaded from the Hugging Face repository allenai/MolmoPoint-8B.

Run MolmoPoint 8B locally

MolmoPoint 8B was made by Ai2 and released on March 18, 2026.

More from Ai2

Yes. MolmoPoint 8B has open weights, so it is free to download and use.