DeepSeek V4.1 Flash: Faster and Cheaper Than V4 Pro
DeepSeek plans to release V4.1 Flash, outperforming V4 Pro at a lower cost, sparking community debate on forced model ...
vLLM v0.29.0 Released: Model Runner V2 Default and More
vLLM v0.29.0 is released with Model Runner V2 as default, FlashInfer all-reduce, new admission control flags, and sup ...
Benchmarking Qwen3.8 27B Quantizations: Performance Analysis
Detailed benchmark results reveal performance changes in Qwen3.8 27B quantizations, showing high performance down to ...
Unsloth v0.1.807-beta Released with AMD and Apple Silicon Boosts
Unsloth v0.1.807-beta is out, bringing major performance boosts for AMD and Apple Silicon, Vulkan defaults, and Windo ...
MiniCPM5-2B Text Generation Model: Our Test Answers, 4GB+ VRAM
OpenBMB has released MiniCPM5-2B, a 2B-class on-device model delivering SOTA performance in coding, math, long contex ...
ExLlamaV3 1.4.8 Released: VRAM and FA2 Updates
ExLlamaV3 1.4.8 is released. Key updates include VRAM allocation optimization and the complete removal of Flash Atten ...