KTransformers
PulseAugur coverage of KTransformers — every cluster mentioning KTransformers across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
XingChen-AGI releases Xing4.0-29B-A4B with 256K context length
XingChen-AGI has released Xing4.0-29B-A4B, a new large language model in the Xing series, formerly known as TeleChat. This model boasts 29 billion parameters with only 4 billion activated per token, enabling a native co…
-
User seeks advice on KTransformers vs llamacpp for MoE optimization
A user on Reddit is seeking advice regarding the performance and inference speed of two different software libraries, KTransformers and llamacpp. The user is specifically interested in optimizing performance for the Qwe…
-
Z.ai releases open-source GLM-5.3-Flash multimodal MoE model
Z.ai has released GLM-5.3-Flash, a new 320-billion-parameter mixture-of-experts model that is natively multimodal and open-source. This model features a hybrid attention architecture combining sparse and linear attentio…
-
FreeToken enables large MoE models to run on single GPUs
Researchers from UC Berkeley and UT Austin have developed FreeToken, an open-source serving engine designed to run large Mixture-of-Experts (MoE) models on single workstation GPUs, significantly reducing the hardware ba…
-
KTransformers updates MoE fine-tuning; Alibaba cuts Qwen3.6 prices
KTransformers has released version 0.7.0, enhancing its capabilities for Mixture of Experts (MoE) fine-tuning. Concurrently, Alibaba Group has reduced the pricing for its Qwen3.6 model, making it more accessible. These …
-
5 LLMs for Local Laptop Coding in 2026
The author highlights five large language models suitable for running locally on a laptop in 2026, emphasizing that smaller, quantized models are now capable of handling significant coding tasks. Qwen2.5-Coder is recomm…
-
Qingjing Technology establishes East China HQ, plans 10k-card AI Token factory
Qingjing Technology, a company specializing in AI Token production services, has established its East China regional headquarters in Qianjiang Century City, Hangzhou. The company plans to build a high-quality AI Token f…
-
DeepSeek V4 Pro hits 40% faster local AI; AlphaGo re-implementation offers LLM insights
An independent developer has optimized DeepSeek V4 Pro for local desktop performance, achieving a 40% speed increase. Concurrently, DeepSeek V4 Flash is now runnable on consumer hardware with 24GB of VRAM using KTransfo…