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 0.46%. The Swift Ornith 9B model reduces thinking token usage by 33.5% without any accuracy loss, and also significantly decreases overthinking loops. Both models are available in various quantizations, including GGUF, and UkisAI is also working on GSQ-RCO implementations. AI
IMPACT These models offer improved efficiency for local LLM deployments, potentially enabling users with less VRAM to run more capable models.
RANK_REASON This is a release of new models by a specific developer, but not from a tier-1 frontier lab.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
