Ai2United States2026

Molmo 2-ER

An AI model: a vision-language-action model (VLA) takes camera images and an instruction and turns them into movements for a robot. This page describes the software, not a robot.

Robotsnot stated

Key figures as stated by the developer

5 of 6 stated · 1 gap

Model type ⁠A embodied reasoning
Controls not stated
Inputs ⁠A camera language
Access ⁠B open download
Parameters ⁠A 4,000,000,000
Base model ⁠A Molmo2-4B (Qwen3-4B backbone + SigLIP2 vision encoder)

How to read the figures

14kgA30.9lbA: The letter in the grey square after a figure points to the source the figure came from. Robots with several sources carry several letters.
74.5lb: Imperial figures are our own conversion unless the manufacturer states an imperial figure itself. The source letter always belongs to the metric figure — a conversion has no source of its own.

The four states

Versions and notes in the key figures (6)

Model type

Developer wording: "Molmo2-ER (Embodied Reasoning) is a 4B vision–language model specialized for the embodied perception skills that downstream action models depend on"; the blog calls it "a specialized embodied-reasoning variant of Molmo 2".

Controls

Note The model answers in text and points, not with robot motion; its skills are "scene understanding, pixel-accurate pointing, multi-image and egocentric–exocentric correspondence, and video temporal reasoning".

Inputs

Model card: the skills are "scene understanding, pixel-accurate pointing, multi-image and egocentric–exocentric correspondence, and video temporal reasoning"; the card's usage example gives an image and a text question.

Access

Paper abstract, which covers the Molmo 2-ER backbone: "We release model weights, training code, and complete training data."

Parameters

Model card: "Molmo2-ER (Embodied Reasoning) is a 4B vision–language model specialized for the embodied perception skills that downstream action models depend on".

Base model

"It is built on top of Molmo2 (Qwen3-4B backbone + SigLIP2 vision encoder)".

Full specifications

6 of 12 fields stated

The letter points to the source at the foot of the page. Fields without a figure are dashed: the developer does not state them — that is not the same as zero, and not the same as no.
Field	Value	Source and caveats
Input and output
Model type	`embodied reasoning`	A Developer wording: "Molmo2-ER (Embodied Reasoning) is a 4B vision–language model specialized for the embodied perception skills that downstream action models depend on"; the blog calls it "a specialized embodied-reasoning variant of Molmo 2".
Inputs	`camera` `language`	A Model card: the skills are "scene understanding, pixel-accurate pointing, multi-image and egocentric–exocentric correspondence, and video temporal reasoning"; the card's usage example gives an image and a text question.
Controls	not stated	Note The model answers in text and points, not with robot motion; its skills are "scene understanding, pixel-accurate pointing, multi-image and egocentric–exocentric correspondence, and video temporal reasoning". ⁠A
Degrees of freedom controlled	not stated
Size and speed
Parameters	4,000,000,000	A Model card: "Molmo2-ER (Embodied Reasoning) is a 4B vision–language model specialized for the embodied perception skills that downstream action models depend on".
Control rate	not stated
Where it runs and what it is built on
Runs where	not stated
Base model	Molmo2-4B (Qwen3-4B backbone + SigLIP2 vision encoder)	A "It is built on top of Molmo2 (Qwen3-4B backbone + SigLIP2 vision encoder)".
Openness
Access	open download	B Paper abstract, which covers the Molmo 2-ER backbone: "We release model weights, training code, and complete training data."
License	Apache-2.0	A Model card: "License: apache-2.0".
The developer’s own claims
Learns new tasks	not stated
Works across robots	not stated	Yes only if the developer itself says the model runs on different robot models, for example robots from two manufacturers or two of one manufacturer’s models. Several units of the same robot do not count.

Notes

Ai2 describes Molmo 2-ER as a 4B vision-language model specialised for the embodied perception skills that action models depend on.
Molmo 2-ER is the backbone of MolmoAct 2 and is also released on its own.

Sources

A Manufacturer product page. huggingface.co/allenai/Molmo2-ER · Retrieved 2026-10-14
B arxiv.org/abs/2605.02881 · Retrieved 2026-10-14

One line per source, not per figure: a source has one retrieval date. The letter beside a figure points to the line it came from.