Introducing Swift: more confident, faster responses

Today, we’re releasing Swift, a new model line from UkisAI. Swift delivers more confident answers, faster responses, and lower compute costs. More confident. Less compute. Faster responses.

More confident. Faster responses.

Swift gives you more decisive answers and reaches them faster. It also takes less compute to produce each response, lowering inference costs.

• More confident answers: clear, decisive responses.
• Less compute: lower inference costs for each response.
• Faster responses: reach useful answers sooner.

Available now

Swift-Qwen3.8-27b is available now on Hugging Face, along with its model card.

A new home for our releases

Our releases now have dedicated Models and Datasets pages. Explore the Swift family in Models, or find our agent-traces release in Datasets.