Multimodal model

Inkling Small

Inkling Small is an open weights Mixture of Experts transformer from Thinking Machines Lab with 276B total parameters and 12B active. It offers native reasoning over text, images, and audio, variable thinking effort, and a context window of up to 1M tokens, achieving performance comparable to the larger Inkling model at a quarter of its size.

Type
Multimodal
Weights
Open
Downloads
623k
Likes
416
Pipeline
image text to text
License
apache-2.0

Model hub

thinkingmachines/Inkling-Small

Repo
thinkingmachines/Inkling-Small
Library
transformers
Access
Public
Technical tags and languages
transformers safetensors inkling_mm_model image-text-to-text conversational audio-text-to-text moe eval-results endpoints_compatible
Open on Hugging Face

More from Thinking Machines Lab

3

About Thinking Machines Lab

1

Thinking Machines Lab develops multimodal AI models, research systems, and model-customization tools.