Skip to content
Latent Space

Thinking Machines' Native Interaction Models: TML-Interaction-Small 276B-A12B Advances Realtime Voice

[AINews] Thinking Machines' Native Interaction Models - TML-Interaction-Small 276B-A12B - advances SOTA Realtime Voice and kills standard VAD

Thinking Machines released TML-Interaction-Small, a 276B-parameter MoE model with 12B active parameters, and the post says it advances realtime voice through 200ms time-aligned microturns, encoder-free early fusion for audio and images under 200ms, and benchmark wins over GPT-Realtime-2 and Gemini 3.1-Flash.

Why it matters: Thinking Machines released TML-Interaction-Small, a 276B-parameter MoE with 12B active, building realtime voice interaction into the model instead of bolting on separate speech recognition and synthesis modules.

Read the original ↗Export Markdown