PulseAugur
EN
LIVE 04:10:06

Rock Matrix open-sources EdgeRock-4B with drastically reduced KV cache

By PulseAugur Editorial · [1 sources] ·

Rock Matrix has open-sourced EdgeRock-4B, a model derived from Qwen3.5-4B. This new model significantly reduces the KV cache size from 8 GiB to 72 MiB while maintaining a 256K context length. EdgeRock-4B achieves over six times the throughput of its predecessor, with minimal impact on its reasoning capabilities. AI

IMPACT This release offers a more efficient model for edge deployments, potentially lowering inference costs and enabling wider use in resource-constrained environments.

RANK_REASON Open-source release of a new model with performance improvements. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Pandaily →

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

Rock Matrix open-sources EdgeRock-4B with drastically reduced KV cache

How we ranked this

Signal score
46 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Open-source release of a new model with performance improvements. [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, infra
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. Pandaily TIER_1 English(EN) · [email protected] (Pandaily) ·

    Rock Matrix Open-Sources EdgeRock-4B, Cutting Qwen3.5-4B's 256K KV Cache From 8 GiB to 72 MiB

    Built from Qwen3.5-4B with HKUST and CMU, EdgeRock-4B keeps a fixed 72 MiB KV cache at 256K context and serves 6.42 times the throughput while keeping most of its reasoning scores.

RELATED ENTITIES

RELATED TOPICS