# H Company's NeoMME rethinks visual document retrieval

> A single-tower multimodal encoder aims to make multilingual document search cheaper to fine-tune and run.

Published by The Open Weights on Aug 30, 2026. Canonical: https://theopenweights.com/news/neomme-f44g

## Key facts

- Company: H company
- Model: NeoMME
- Version: NeoMME
- Category: Embeddings
- Modalities: Embeddings, Vision-Language
- License: Other (Open weights, commercial use allowed)
- Significance: notable
- Published: 2026-08-30
- Last verified: 2026-09-05
- Canonical URL: https://theopenweights.com/news/neomme-f44g

H Company has introduced **NeoMME**, a foundation encoder built specifically for visual document retrieval — the task of finding relevant pages, screenshots, and richly formatted documents rather than plain text passages. According to the company's [write-up on Hugging Face](https://huggingface.co/blog/Hcompany/neomme), the model is designed to be multimodal-native and multilingual from the ground up, so images and text are handled by one shared pathway.

The headline architectural choice is a **single-tower** design. Many retrieval systems use separate encoders for queries and documents, which adds complexity and cost. Folding both into one tower is meant to simplify training and deployment while keeping the model versatile across languages and document types.

## Why it matters

Retrieval-augmented systems increasingly need to search over what documents actually look like — tables, charts, forms, and scanned pages — not just their extracted text. NeoMME targets that gap with a few practical priorities:

- **Compact late-interaction embeddings**, which retain fine-grained matching between query and document tokens without ballooning storage.
- **Multilingual coverage**, so the same encoder can serve document search across languages.
- **Efficient fine-tuning and inference**, positioned as a key advantage for teams operating at scale.

Late-interaction methods, popularized by ColBERT-style retrieval, tend to be more accurate than single-vector approaches but heavier to store and serve. By emphasizing compact embeddings, NeoMME is pitching a balance between retrieval quality and operational cost.

The release is described as an initial foundation encoder under a non-standard license, so teams will want to check terms before production use. Full benchmark details and training specifics are laid out in the accompanying [paper](https://huggingface.co/papers/2609.01657).

## Get the model

- [HF Papers](https://huggingface.co/papers/2609.01657)

## Sources

- [NeoMME: A Single-Tower Multimodal-Native Multilingual Foundation Encoder for Efficient Fine-Tuning and Inference](https://huggingface.co/papers/2609.01657) — HF Papers, Aug 30, 2026

---
Source: The Open Weights (https://theopenweights.com/). Aggregated and written by Claude, curated by humans. Cite as: "H Company's NeoMME rethinks visual document retrieval", The Open Weights, Aug 30, 2026, https://theopenweights.com/news/neomme-f44g