Aikido Security Releases Altar-1: An Open-Weight Security Model Pruned From GLM-5.3 to 328 GB
2 sources 4 min
Qualcomm launches two new smartphone chips with emphasis on AI
2 sources 2 min
Hunyuan-A13B Technical Report
[AINews] Xiaomi MiMo-V2.6-Pro 1T-A42B: the new top Open Weights model, trained for $3M
5 sources 13 min
OPBackdoor: Opportunistic Backdoors via Alibi-Aligned Reasoning
StepFun Launches Step 5 Preview: A 600B-Total, 27B-Active MoE Model With 1M Context for Long-Horizon Agentic Work
4 min
All-in-One Multilingual Scene Text Recognition with Script-aware Mixture-of-Experts
Jina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUs
5 min1
IntBMoE: Integrating Block-Level Conditioning into Expert Composition for Full-Participation Mixture-of-Experts
DeepSeek-v4.1 Flash: Pushing the Limits of KV Cache Compression
15 min1
Higher-order pruning of experts in mixture-of-experts language models
1
XingChen-AGI/Xing4.0-29B-A4B — new model trending #30 on Hugging Face
7 min1
MoME: Mixture-of-Memory Embeddings for Context-Aware Sparse Lookup
2
[AINews] DeepSeek v4.1-Flash: 763B-P8B-D16B novel causal Encoder–Decoder architecture with vision marks the Return of the Whale
15 min1
Expert-Space Exploration in MoE Reinforcement Learning
Cohere Releases North Small Translate: A 218B MoE Translation Model That Scores 83.6 on WMT26 Across 50 Languages
4 min
Expert-Space Exploration in MoE Reinforcement Learning
Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data
New Deepseek model V4.1-Flash cuts memory needs for AI agents
4 min1
deepseek-ai/DeepSeek-V4.1-Flash — new model trending #28 on Hugging Face
10 min1
T1: Terminal Agent Reinforcement Learning for Long-Horizon Tasks
1
Pelican-Sim 1.0: A General World Model Simulator for Embodied Intelligence
nex-agi/Nex-N2.5-mini — new model trending #30 on Hugging Face
13 min1