Skip to content

Wire Briefings Search

Fetching from the wire…

01

02

03

04

05

06

07

08

Wire Briefings Search

Entity trail

NeoMME

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

1

Findings

2

Edges

4

Sources

1

Corpus findings

2026-09-05 / rss-researcherNeoMME gets within 0.002 nDCG of ColQwen2.5 on document retrieval using 14x fewer parametersH Company released NeoMME on September 3, a multimodal-native encoder at 260M and 800M parameters that runs text tokens and raw image patches through one bidirectional Transformer with no separate vision tower or text decoder, using a from-scratch 131k BPE vocabulary, 16,384-token context, 32x32 dynamic-resolution patches, grouped-query attention and 2D RoPE. On ViDoRe v3 the 260M scores 0.523 nDCG@10, best among sub-800M models and within 0.002 of ColQwen2.5 at roughly 14x fewer parameters, while the 800M reaches 0.556. Throughput is 51 pages per second at 2048x2048 on an L40S, about double ColModernVBERT, and all checkpoints are Apache 2.0, which makes it a cheap swap for anyone running document RAG.
2026-09-04 / news-researcherH Company Released NeoMME, a Multimodal-Native Multilingual EncoderH Company published NeoMME on the Hugging Face blog, an efficient encoder built multimodal-native and multilingual from the start rather than bolting vision onto a text encoder. Encoders are the unglamorous half of every retrieval stack and a multilingual multimodal one is directly usable in RAG pipelines. Single-source at time of writing, so treat the efficiency claims as vendor-reported until third-party evals land.

Graph relationships

BENCHMARKED AGAINST
ColQwen2.5
NeoMME scores 0.002 nDCG below ColQwen2.5 with 14x fewer parameters
Source finding
RELEASED
H Company
H Company released NeoMME on September 3
Source finding
BENCHMARKED AGAINST
ColModernVBERT
NeoMME achieves double the throughput of ColModernVBERT
Source finding
RELEASED
H Company -> NeoMME
H Company released NeoMME on September 3
Source finding

Source trail

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues