startlux-models/StartLux-Decision-v1-2B

VISIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2.3BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Sep 29, 2026License:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

StartLux-Decision-v1-2B by startlux-models is a 2.3 billion parameter model designed for structured decision-making tasks, providing probabilistic answers to typed questions (choice, yes/no, rating) about a given state. It supports multimodal inputs including text, JSON, and images, with a large context window of up to 262,144 tokens. Optimized for fast inference, it processes requests efficiently, making it suitable for applications requiring rapid, probabilistic decision outputs.

Loading preview...

StartLux-Decision-v1-2B Overview

StartLux-Decision-v1-2B is a 2.3 billion parameter model developed by startlux-models, specifically engineered for structured decision-making. It processes typed questions (e.g., multiple choice, yes/no, rating scales) about a given state and returns a probability for each possible option. This model is part of a family of StartLux-Decision models, ranging from 0.8B to 35B parameters.

Key Capabilities

  • Probabilistic Decision Outputs: Provides a probability distribution for each option in response to a question, enabling nuanced decision support.
  • Multimodal Input Support: Accepts diverse input formats including text, JSON, and images, allowing for comprehensive state representation.
  • Extended Context Window: Handles long inputs up to 262,144 tokens (256K), processing complex states efficiently.
  • Fast Inference: Optimized for low-latency inference, achieving 15.5 ms for a request with three questions and 9.6 ms for a single yes/no question on an H200 GPU, utilizing custom kernels and CUDA graphs.
  • TypeSafe API Compatibility: Uses the TypeSafe /v1/systemone format, ensuring seamless integration with existing Jev clients.
  • Batch Processing: Supports decide_batch for efficient processing of multiple requests simultaneously.

Good For

  • Applications requiring rapid, probabilistic answers to structured questions.
  • Systems that need to analyze multimodal data (text, JSON, images) for decision support.
  • Use cases where a large context window is crucial for understanding complex scenarios.
  • Integrating with existing TypeSafe API-compatible infrastructures.
  • Scenarios demanding high throughput and low latency for decision inference.

Finetunes
4 models