PulseAugur
EN
LIVE 21:56:13

H Company releases NeoMME multimodal encoders with single-tower architecture

By PulseAugur Editorial · [1 sources] ·

H Company has introduced NeoMME, a new family of multimodal encoders designed for efficient document retrieval. These models, available in 260M and 800M parameter sizes, uniquely integrate text and image processing within a single transformer tower, eliminating the need for separate vision towers and causal decoders. The NeoMME-Retriever variant demonstrates competitive performance on benchmarks like ViDoRe v3, achieving strong results with significantly fewer parameters than other leading models. The checkpoints are released under Apache 2.0 and are compatible with Hugging Face Transformers, enabling practical deployment. AI

IMPACT Offers a more parameter-efficient approach to multimodal retrieval, potentially lowering deployment costs and enabling wider adoption.

RANK_REASON Release of a new family of multimodal encoders with a novel single-tower architecture and competitive benchmark results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on MarkTechPost →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

H Company releases NeoMME multimodal encoders with single-tower architecture

How we ranked this

Signal score
53 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Release of a new family of multimodal encoders with a novel single-tower architecture and competitive benchmark results. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. MarkTechPost TIER_1 English(EN) · Asif Razzaq ·

    H Company Releases NeoMME: A Family of 260M and 800M Single-Tower Multimodal Encoders That Drop the Vision Tower and Causal Decoder

    <p>We look at NeoMME, a family of 260M and 800M bidirectional encoders from H Company. Unlike ColPali-style retrievers, it processes multilingual text tokens and raw 32×32 image patches in a single Transformer, with no pretrained vision tower and no causal decoder. We cover the m…

RELATED ENTITIES

RELATED TOPICS