Latest news, tutorials, and insights about AI
Liquid AI introduces LFM3, a milestone 2B-class multimodal family featuring a tri-hybrid architecture and six specialized model instances on a shared backbone.
OpenAI's new GPT-6.1 Terra offers a massive leap in computer-use capabilities and visual reasoning at a 50% lower price point than its predecessors.
GrootTeam introduces a new family of four Mixture-of-Experts reasoning models, ranging from 19B to 280B parameters, optimized for high-efficiency complex reasoning.
Moonshot AI unveils Kimi K3.1 Pro, a 2.8T parameter Mixture-of-Experts model featuring 1M context and industry-leading terminal and coding benchmarks.
iFlytek introduces Spark-X2.6-4B, a 4B parameter model optimized for on-device agentic workflows, featuring controllable thinking budgets and Apache-2.0 licensing.
MiniMax officially launches M3.1 Flash, a 428B MoE model optimized for agentic workflows, coding, and computer use with 3x faster inference.
DeepSeek launches V4.1-Flash-1017, an open-weights MoE model delivering massive performance gains in agentic coding and multimodal reasoning while maintaining existing pricing.
NVIDIA quietly drops Dory 1.0, a 4B parameter model featuring a novel looped latent-recurrent architecture for elastic compute deployment.
Deep Reinforce AI launches Ornith-1.5-122B, a high-performance MoE coding model designed for single-GPU deployment and edge-mobile use.
