4 results for “coding benchmarks”
Z.ai released GLM-5.3-Flash, a 320B-parameter multimodal model with only 18B active parameters using hybrid sparse-linear attention. The MIT-licensed model approaches Claude Opus 4.8 on coding benchmarks at one-tenth the inference cost of its predecessor.
Z.ai confirmed Tuesday it developed Ox Alpha, the anonymous open-weight model that topped coding benchmarks after launching on OpenRouter last Thursday. The lab will release model weights Wednesday, giving developers a production-grade reasoning alternative to closed frontier systems.
Meta has entered the AI coding agent arena with Muse Code, a terminal-based tool now in beta that targets professional developers working across large. Meta's entry reshapes the competitive dynamics of AI-assisted software development.
Liquid AI has released the LFM2.5 model family, a comprehensive suite of open-weight models designed to run entirely on edge devices without cloud. The LFM2.5 releases address three critical constraints that have limited on-device AI.