UkisAI
PulseAugur coverage of UkisAI — every cluster mentioning UkisAI across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
UkisAI releases Swift GLM-5.3 and Ornith 9B models with reduced token usage
UkisAI has released two new Swift models: Swift GLM-5.3-Flash and Swift Ornith 9B. The Swift GLM-5.3-Flash model achieves a 61.4% reduction in thinking tokens on the MMLU-PRO benchmark with a minimal accuracy loss of 0.…
-
Swift Qwen 3.8 27B model surpasses 100k downloads, boosts efficiency
UkisAI's Swift Qwen 3.8 27B model has achieved over 100,000 downloads and ranks highly on HuggingFace's trending models. This open-source release demonstrates a method to reduce token usage by 58.3% and increase speed b…
-
UkisAI fine-tunes Qwen3.8-27B to cut reasoning tokens by 40%
UkisAI has released Swift-Qwen3.8-27B, a fine-tuned version of the Qwen3.8-27B model that significantly reduces token usage for reasoning tasks. This new model addresses the 'overthinking' issue present in earlier Qwen …
-
UkisAI optimizes Qwen 3.8 27B for speed and efficiency · 1 source tracked
UkisAI has developed a post-trained version of the Qwen 3.8 27B model, named Swift-Qwen3.8-27B, which significantly reduces "thinking" tokens by 58.3% and increases speed by 1.95x, all while maintaining accuracy with le…