Model Terminal

MolmoAct2-SO101-Mdrift_vtptoponly60-D6231-step42000

A community-uploaded training checkpoint derived from Ai2's MolmoAct2 vision-language-action (VLA) robotics model family, fine-tuned for the SO-100/101 robot platform. It converts image observations, natural-language instructions, and robot state into continuous manipulation actions for real-world robot control. Source: supabase.

Value score

Not scored

Context

—

tokens

Max output

—

tokens

Price

—

Capability radar

Peer value bars

Claude 4 Opus95
GPT-594
Gemini 2.5 Pro93
o393
Claude 4 Sonnet91

Identity

Developer: Unknown
Openness: —
Modalities: —
Release: —
Knowledge cutoff: —
Deprecation: —
API docs: —

Benchmarks

No benchmark scores yet.

Compare nearby

Claude 4 Opus Compare
GPT-5 Compare
Gemini 2.5 Pro Compare
o3 Compare
Claude 4 Sonnet Compare

Pricing

Input / 1M: —
Output / 1M: —
Speed: —