← all releases
OpenBMB·Feb 11, 2026·v1.0·open source

MiniCPM-SALA

MiniCPM-SALA is a large-scale hybrid language model developed by OpenBMB, integrating sparse and linear attention mechanisms to efficiently handle long-context modeling. Released on February 11, 2026, it is the first model to effectively combine sparse and linear attention for million-token context modeling. ([huggingface.co](https://huggingface.co/openbmb/MiniCPM-SALA?utm_source=openai))

LLMsML
visit website
overview

MiniCPM-SALA is a large-scale hybrid language model developed by OpenBMB, integrating sparse and linear attention mechanisms to efficiently handle long-context modeling. Released on February 11, 2026, it is the first model to effectively combine sparse and linear attention for million-token context modeling. ([huggingface.co](https://huggingface.co/openbmb/MiniCPM-SALA?utm_source=openai))

key features
  • 01Hybrid architecture combining 25% sparse attention and 75% linear attention
  • 02Achieves 3.5× inference speedup over dense baselines
  • 03Supports context lengths up to 1 million tokens
  • 04Utilizes Hybrid Positional Encoding (HyPE) for improved performance
  • 05HALO adaptation for efficient transfer of dense attention capabilities
related productsbrowse all →
Step-Audio-R1open sourceStep-Audio-R1 is an advanced audio language model developed by StepFun AI, designed to enhance audio reasoning capabilities by grounding its reasoning in acoustic features. It introduces Modality-Grounded Reasoning Distillation (MGRD), an iterative training framework that shifts the model's reasoning from textual abstractions to acoustic properties, effectively addressing the 'inverted scaling' problem where performance degrades with longer reasoning. This model has demonstrated superior performance across various audio understanding and reasoning benchmarks, surpassing models like Gemini 2.5 Pro and achieving results comparable to Gemini 3 Pro.SINQopen sourceSINQ (Sinkhorn-Normalized Quantization) is a novel, fast, and high-quality quantization method designed to make any Large Language Model (LLM) smaller while preserving accuracy. It offers a plug-and-play, model-agnostic technique that delivers state-of-the-art performance for LLMs without sacrificing accuracy.AIBuildAIopen sourceAIBuildAI is an AI agent that autonomously constructs AI models. Given a specific task, it initiates an agent loop to analyze the problem, design models, and execute training processes, all without human intervention.