Latest

Every published item, in order · 9305 items
Nov 2
Tomorrow 08:00, the daily lands in your inboxTelegram channel or email. Same screening as this page, no ads.
Nov 1
ModelsBuildersSelected15:12
Rock Matrix Open-Sources EdgeRock-4B Reasoning Model With Fixed Memory

A fixed-size KV cache means long-horizon reasoning on consumer GPUs without the memory creep that normally forces context truncation, making the 4B model a practical candidate for on-device and edge deployments. Developers building local agents or long-session assistants can run it without enterprise hardware. Compared with other small open models, its benchmark is framed against bounded-state baselines rather than full-cache models, so independent verification will matter.

TestingCatalog@testingcatalogSource
ModelsBuildersMulti-sourceSelected09:07
SenseNova Loop-9B-Preview open-source model matches 27B rivals

Parameter-sharing loops let a 9B model punch at 27B weight class, which cuts inference cost for teams running complex agent pipelines on limited hardware. Developers building coding or office-automation agents can download the weights from Hugging Face and benchmark them against Qwen3.5-27B directly. The approach also signals a broader shift in Chinese labs toward compute-depth efficiency over raw parameter scaling.

Event
IT之家Source
Oct 31
Oct 30
ModelsBuildersMulti-sourceSelected21:18
Xiaomi open-sources 0.8B OCR model Xiaomi-OCR-0

A sub-1B model reaching near-top OmniDocBench scores means document parsing and invoice extraction can now run on modest hardware rather than large multimodal models. Teams building document pipelines can deploy it locally, pull weights from Hugging Face or GitHub, and use the JSON and Markdown outputs directly in automation workflows. Against rivals like PaddleOCR's larger document models, it trades scale for efficiency while matching them on benchmarks.

Event
IT之家Source

Only the latest 2000 items are listed; for older ones see the daily archive