LinkSubquadratic Inc.Announcement
Scaling standard Transformer models to millions of tokens has historically been hindered by quadratic compute costs and context degradation. If subquadratic approaches like SSA prove reliable, enterprises can process entire code repositories or documentation archives in a single context window without the complexity of RAG pipelines.
Subquadratic targets fundamental architectural constraints rather than incremental optimization. By shifting compute focus only to significant token interactions, SubQ aims to make massive context windows computationally viable for agentic workflows and repository-scale reasoning. The real-world utility of these gains will become clearer as broader API access rolls out beyond early access testing.
Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.
Similar in content, found by the AI search.
Article:Kimi Team (Yu Zhang et al.)
The Kimi Team presents Kimi Linear, a hybrid linear attention architecture that outperforms standard full attention across short-context, long-context, and RL scaling scenarios. It cuts KV cache consumption by up to 75% and boosts decoding throughput up to sixfold on 1M token contexts.
Article:Albert Gu, Tri Dao
Albert Gu and Tri Dao introduce Mamba, a sequence modeling architecture based on selective structured state space models (SSMs). It achieves linear scaling with sequence length, eliminates attention and MLP blocks, and delivers five times higher inference throughput than Transformers.
Link:Z.ai
Z.ai has released GLM-5.2 on Hugging Face under the permissive MIT licence. The flagship open-weights model is designed for long-horizon agentic and coding workflows, featuring a robust 1M-token context window. Architectural updates including IndexShare and an enhanced MTP layer significantly cut computation FLOPs and accelerate speculative decoding.
Link:Qwen
The Qwen team has released Qwen3.8-27B, a compact, dense vision-language model under the Apache-2.0 licence. The 27-billion-parameter model pairs Gated DeltaNet with conventional gated attention, handles images and videos natively, and introduces flexible reasoning controls.
Link:Together AI
Together AI has introduced Qwen's flagship proprietary model, Qwen3.7-Max, to its API platform. Built for the agent era, the text-only model features a 1M token context window and focuses on software engineering, office automation, and long-horizon task execution.
Link:Apodex AI
Apodex AI has released Apodex-1.1-mini under the Apache-2.0 licence on Hugging Face. Based on the qwen3_5_moe architecture, the model is designed for long-horizon research and agentic workflows with native tool execution.
Parts of this content were created with the assistance of AI.
