MolmoWeb

Model ● Covered in 4 stories + Follow

MolmoWeb is an open-source visual web agent built on the Molmo 2 multimodal model that navigates websites and completes browser tasks by interpreting screenshots. Available in 4B and 8B parameter sizes, the 8B variant achieves 78.2% performance on the WebVoyager benchmark and outperforms other open-weight agents. The model has been integrated into accessibility testing tools and web automation applications.

Updated 8 August 2026

Specifications

No specifications recorded yet.

Latest developments

Building accessibility tools on a truly open foundation Allen Institute (AI2) · 3 months ago · 51

Molmo learns to point and act Allen Institute (AI2) · 4 months ago · 41

Deep Learning Weekly: Issue 448 Deep Learning Weekly · 5 months ago · 45

MolmoWeb: An open agent for automating web tasks Allen Institute (AI2) · 5 months ago · 55

Timeline

Month Quarter Year

Q2 2026

21 May Building accessibility tools on a truly open foundation
29 Apr Molmo learns to point and act

Q1 2026

Google Research Introduces TurboQuant, a Vector Compression Algorithm for Key-Value Cache Optimization Research publication

26 Mar Deep Learning Weekly: Issue 448
24 Mar MolmoWeb: An open agent for automating web tasks

Relationships

Products & technology

Derived from Molmo 2 · 1 source
Derived from Molmo · 1 source
Molmo develops this model · 1 source
AI2 develops this model · 1 source
PointCheck integrated with this model · 1 source

Every AI story that matters,
in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.