MiniMax M2.5 Uses 10B Active Parameters per Token, Aiming for Cheaper Always-On Agents

MiniMax M2.5 hits about 80% on Sweetbench and runs near 100 tokens per second, helping teams deploy faster models on tighter budgets.
IEEE

A Hardware-Efficient Sparse-Matrix Multiplier for Deep Learning Models Using CSR/CSC

Abstract: Deep learning models rely heavily on matrix multiplication, which is computationally expensive and memory-intensive. Sparse matrices, which contain a high proportion of zero elements, offer ...
IEEE

MH-SpGEMM: Efficient Sparse General Matrix-Matrix Multiplication on Modern GPUs via Masking and Hashing Cooperative Optimization

Abstract: Sparse General Matrix-Matrix Multiplication (SpGEMM) is a core operation in high-performance computing applications such as algebraic multigrid solvers, machine learning, and graph ...
MarketWatch

Matrix Applications Completes SOC 1 and SOC 2 Type 2 for TradeBlazer Collateral Management System and Managed Services

The MarketWatch News Department was not involved in the creation of this content. New York, NY, Jan. 20, 2026 (GLOBE NEWSWIRE) -- Matrix Applications, LLC ("Matrix") has successfully completed the ...
Business Insider

Matrix Applications Completes SOC 1 and SOC 2 Type 2 for TradeBlazer Collateral Management System and Managed Services

New York, NY, Jan. 20, 2026 (GLOBE NEWSWIRE) -- Matrix Applications, LLC ("Matrix") has successfully completed the 2025 System and Organization Control (SOC) 1 and SOC 2 Type 2 audits for its ...