Shanghai AI startup StartLux explores opportunities in decision models and local AI

As artificial intelligence moves from content generation to decision-making and autonomous task execution, specialized decision models and locally deployed AI are opening up new commercial opportunities. Shanghai-based startup StartLux is exploring both areas.

StartLux-Decision v1.1 27B Instant recently ranked first among 174 entries on Hugging Face — a United States-based platform for sharing AI models and datasets — in its Decision Index 0.3.2 leaderboard with a score of 69.12, according to company-published results.

Chuan-Wei Kuo, co-founder and chief technology officer of StartLux, said decision models could reduce computing costs and improve efficiency. Unlike general-purpose large language models which primarily generate text, decision models select from predefined options and return probabilities. This makes them suitable for customer service routing, web operations, and AI agent execution.

Gao Chengyuan, head of Influence Academy, said decision quality and speed were becoming increasingly important as AI moved from "talking" to "doing". Faster decisions could shorten the process from perception and judgment to execution, improving responsiveness and efficiency.

Kuo expects more than 80 percent of AI demand to be handled locally within the next one to two years. StartLux plans to launch a consumer-facing local AI agent, followed by enterprise solutions next year. The products are designed to process data locally rather than send it to third-party servers, giving users greater control over privacy and data.

Local deployment could be particularly relevant to government, financial, and industrial customers that require data security, low latency, and cost control, Gao said.

Competition in decision models is intensifying, with companies and research institutions including OpenAI, Cloudflare, Amazon, and Shanghai Artificial Intelligence Laboratory introducing related models or research initiatives. Kuo said the field remains exploratory, with progress depending on technological breakthroughs and practical applications.

Looking ahead, Kuo said StartLux will continue developing its recursive self-improvement framework. Its Auto Research approach uses AI to generate hypotheses, conduct experiments, and evaluate results, aiming to make AI research and development more efficient over time.