Qwen3.8-Flash-Next Multimodal MoE Model: ~402GB Memory
Explore Qwen3.8-Flash-Next, a next-gen open-weight multimodal causal language model featuring hybrid attention, MoE a ...
SGLang v0.5.21 Released with Dynamic PD Role Switching
SGLang v0.5.21 is out, bringing dynamic PD role switching, a Rust-based prefix cache core, new APIs, and expanded mod ...
Xiaomi Releases MiMo-V2.6-Flash-MOPD and Pro-MOPD Models
Explore Xiaomi’s new sparse MoE multimodal models featuring MOPD2 distillation, 1M context, and improved agent ...
DeepSeek-V4-Pro-0813 Text Generation Model: ~998GB Memory, GGUF Builds
Discover DeepSeek-V4-Pro-0813, a 1.6T MoE model on Hugging Face featuring enhanced agent capabilities, 1M context, an ...
LFM2.5-VL-3B-DSpark Draft Model for Vision-Language Models: 4GB+ VRAM
LiquidAI has released LFM2.5-VL-DSpark, an experimental DSpark draft model for LFM2.5-VL-3B, speeding up inference wi ...
K2-Horizon-32B-NVFP4: Our Test Answers, 32GB+ VRAM
IFM releases K2-Horizon-32B-NVFP4, an NVFP4 quantized version of the 32B long-context open-weight LLM optimized for N ...
K2-Horizon-375B-A23B-NVFP4 Text Generation Model: ~257GB Memory
Explore the NVFP4 quantized version of IFM’s K2-Horizon-375B-A23B MoE model, designed for NVIDIA Blackwell GPUs ...
SGLang v0.5.20 Released: New Models and Optimizations
SGLang v0.5.20 is released with 713 PRs, adding support for new models, ROCm load time improvements, Intel XPU offici ...
Simple-Attention-Sparsification Text Generation Model: 12GB+ VRAM
Tencent has released Simple Attention Sparsification (SAS) checkpoints for Qwen3 models, enabling efficient context p ...
DeepSeek-V4-Pro-0813-nvfp4-DSpark: ~1005GB Memory
NVIDIA has released nvidia/DeepSeek-V4-Pro-0813-nvfp4-DSpark, a fully NVFP4-quantized MoE language model supporting S ...