---
license: apache-2.0
datasets:
- allenai/MolmoWeb-SyntheticTraj
- allenai/MolmoWeb-HumanTrajs
- allenai/MolmoWeb-HumanSkills
- allenai/MolmoWeb-SyntheticSkills
- allenai/MolmoWeb-SyntheticQA
- allenai/MolmoWeb-SyntheticGround
language:
- en
base_model:
- Qwen/Qwen3-8B
- google/siglip-so400m-patch14-384
pipeline_tag: image-text-to-text
library_name: transformers
tags:
- multimodal
- olmo
- molmo
- molmo2
---

<img src="molmoweb_logo.png" alt="Logo for the MolmoWeb Project" style="width: auto; height: 50px;">

# MolmoWeb-8B

<span style="color:red; font-weight: bold;">Important Update!</span> 
We made a few small but important updates to this HF/transformers-compatible checkpoint to ensure exact outputs to our native model checkpoint on **March 29, 2026 ~6PM PST**. 
If you downloaded this model checkpoint earlier than this time, we recommend re-downloading it. See PRs [2](https://huggingface.co/allenai/MolmoWeb-8B/discussions/2) and [3](https://huggingface.co/allenai/MolmoWeb-8B/discussions/3) for more details. Thanks for your understanding! 

MolmoWeb is a family of fully open multimodal web agents. MolmoWeb agents achieve state-of-the-art results outperforming similar scale open-weight-only
models such as Fara-7B, UI-Tars-1.5-7B, and Holo1-7B. MolmoWeb-8B also surpasses set-of-marks
(SoM) agents built on much larger closed frontier models like GPT-4o. We further demonstrate
consistent gains through test-time scaling via parallel rollouts with best-of-N selection, achieving 94.7%
and 60.5% pass@4 (compared to 78.2% and 35.3% pass@1)on WebVoyager and Online-Mind2Web
respectively. 

**Learn more** about the MolmoWeb family in our announcement [blog post](https://allenai.org/blog/molmoweb) and [tech report](https://allenai.org/papers/molmoweb).

MolmoWeb-8B is based on [Molmo2](https://arxiv.org/abs/2601.10611) architecture, which uses [Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) and [SigLIP 2](https://huggingface.co/google/siglip-so400m-patch14-384) as vision backbone. 

Ai2 is committed to open science. The MolmoWeb datasets are available [here](https://huggingface.co/collections/allenai/molmoweb-data). 
All other artifacts used in creating MolmoWeb (training code, [evaluations](https://github.com/allenai/molmoweb), intermediate checkpoints) will be made available, furthering our commitment to open-source AI development and reproducibility.

Quick links:
- 💬 [Demo](https://molmoweb.allen.ai/)
- 📂 [All Models](https://huggingface.co/collections/allenai/molmoweb)
- 📚 [All Data](https://huggingface.co/collections/allenai/molmoweb-data)
- 📃 [Paper](https://allenai.org/papers/molmoweb)
- 🎥 [Blog with Videos](https://allenai.org/blog/molmoweb)

## Quick Start
```python
from transformers import AutoProcessor, AutoModelForImageTextToText
from PIL import Image
import requests
import torch
from jinja2 import Template

checkpoint_dir = "allenai/MolmoWeb-8B"

model = AutoModelForImageTextToText.from_pretrained(
    checkpoint_dir,
    trust_remote_code=True,
    torch_dtype=torch.float32, # we recommend using the default float32 precision 
    attn_implementation="sdpa",
    device_map="auto",
)

processor = AutoProcessor.from_pretrained(
    checkpoint_dir,
    trust_remote_code=True,
    padding_side="left",
)


MOLMOWEB_THINK_TEMPLATE = Template(
"""
# GOAL
{{ task_description }}

# PREVIOUS STEPS
{% for action in past_actions: -%}
## Step {{ action['index'] }}
THOUGHT: {{ action['thought'] }}
ACTION: {{ action['action'] }}
{% endfor %}
# CURRENTLY ACTIVE PAGE
Page {{ page_index }}: {{ page_title }} | {{ page_url }}

# NEXT STEP

"""
)

task_description = "Tell me about the Ai2 PIROR team's recent projects"
past_actions = []
user_message = MOLMOWEB_THINK_TEMPLATE.render(
    page_title=None,
    page_url="about:blank",
    page_index=0,
    task_description=task_description,
    past_actions=[]
)
system_message = "molmo_web_think"
prompt = f"{system_message}: {user_message}"

blank_image = Image.new("RGB", (1280, 720), color="white")

image_messages = [
    {
        "role": "user",
        "content": [
            {"type": "text", "text": prompt},
            {"type": "image", "image": blank_image},
        ]
    }
]

inputs = processor.apply_chat_template(
    image_messages,
    tokenize=True,
    add_generation_prompt=True,
    return_tensors="pt",
    return_dict=True,
    padding=True,
)

# Remove token_type_ids: HF uses it to enable bidirectional attention for image tokens; molmoweb is trained with causal attention only
inputs = {k: v.to("cuda") for k, v in inputs.items() if k != "token_type_ids"} 

with torch.inference_mode():
    output = model.generate(**inputs, max_new_tokens=200)

generated_tokens = output[0, inputs["input_ids"].size(1):]
print(processor.decode(generated_tokens, skip_special_tokens=True))
```

## License and Use

This model is licensed under Apache 2.0. It is intended for research and educational use in accordance with Ai2’s [Responsible Use Guidelines](https://allenai.org/responsible-use).