FastVLM 1.5B

Estimated VRAM is 1.8 GB at Q4, 2.8 GB at Q8 and 4.9 GB at FP16, based on the 1.9B parameter count. These are estimates: weights plus about 15% and 0.5 GB for a short context.

Yes. FastVLM 1.5B fits on a 24 GB, 16 GB, 12 GB and 8 GB GPU at FP16, with an estimated 4.9 GB of VRAM.

Yes. At Q4 the estimated VRAM is 1.8 GB, and assuming macOS lets the GPU use about 75% of unified memory, the smallest common Mac memory size that fits Q4 is an estimated 8 GB.

Yes. FastVLM 1.5B has open weights, so you can self-host it for full control of your data.

Run FastVLM 1.5B locally

No audio, video or PDF support. FastVLM 1.5B accepts text and image inputs only.

Matched on model kind and release date (no comparable price): Llama Prompt Guard 2 22M by Meta, Llama Prompt Guard 2 86M by Meta and Qwen3 8B by Alibaba.

Llama Prompt Guard 2 22MFastVLM 1.5B vs Llama Prompt Guard 2 22MLlama Prompt Guard 2 86MFastVLM 1.5B vs Llama Prompt Guard 2 86MQwen3 8BFastVLM 1.5B vs Qwen3 8B

Open weights. FastVLM 1.5B is released under the apple-amlr license.

The weights are available in the Hugging Face repository apple/FastVLM-1.5B.

Run FastVLM 1.5B locally

Apple, released on May 7, 2025.

More from Apple

FastVLM 1.5B has open weights, so you can download and self-host it.