Skip to content

MM1

A 30B multimodal LLM from Apple with SFT tuning

Visit official site
Category AI models

What MM1 is

The model is described as an Apple multimodal LLM with sizes up to 30B, tuned through pre-training and SFT.

What you can do with it

  • Work on image captioning
  • Work on visual question answering
  • Study in-context prediction

Benchmark notes

  • Results are reported across several benchmarks
  • SFT is described as part of the training
  • Multi-image reasoning is listed as a property

Pros & cons

✓ What we like

  • In-context prediction is named as a property
  • Multi-image reasoning is named as a property

! What to watch out for

  • Results are reported on the named benchmarks
  • The material describes SFT as part of the method

FAQ

How large is the model?

The material states sizes up to 30B.

Which abilities are listed?

In-context prediction, multi-image reasoning and few-shot learning are named.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review