StartLux

Headlines · 2

1. Oct 23
   StartLux open-sources Prism inference engine for decision models

   StartLux has open-sourced StartLux-Prism, an inference engine designed for StartLux-Decision. According to StartLux’s tests on one NVIDIA H200, Prism ran 1.2 to 5.3 times faster than the original vLLM setup depending on the workload, while supporting combined tensor and data parallelism and deterministic output in supported configurations. The reported performance figures are vendor-published and vary with request pattern, shared-context reuse, input length, concurrency and hardware configuration.

   InfoQ 中文 AI&大模型 · 🔥 0
2. Oct 18
   StartLux-Decision-v1.1 tops Hugging Face decision-model leaderboard

   Reports from 36Kr, TMTPost and InfoQ Chinese say StartLux's StartLux-Decision-v1.1-27B Instant ranked first on Hugging Face's updated Decision Index 0.3.2 with 69.1 points, which InfoQ gives as 69.12; the reports differ on whether 173 or 174 entries were listed. The new generation offers Instant, which returns option probabilities in a single pass, and AutoThink, which generates extra reasoning only when initial confidence is insufficient; the team also introduced the RLCD² training method and the DecisionPivot benchmark, and says 27B AutoThink scored 73.37 on its own Decision Index 0.3 evaluation versus 57.96 for hosted Jev. Model weights and inference code are published, but the weights use a CC BY-NC 4.0 license limited to non-commercial use, with commercial deployment requiring separate authorization.

   InfoQ 中文 AI&大模型 · 🔥 0

Experience and discussion from the community

Share my StartLux experienceAsk about StartLux