nics-efc/Standard-1.7B

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Nov 23, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The nics-efc/Standard-1.7B is a 2 billion parameter language model developed by nics-efc, designed for general applications. It is specifically trained on a diverse mixture of math, code, and science data, making it well-suited for reasoning tasks across these domains. This model is presented in the paper "Think-at-Hard: Selective Latent Iterations to Improve Reasoning Language Models," highlighting its focus on enhancing reasoning capabilities. With a 32768 token context length, it can process substantial amounts of information for complex problem-solving.

Loading preview...

Model Overview

The nics-efc/Standard-1.7B is a 2 billion parameter general-purpose language model developed by nics-efc. It is distinguished by its specialized training regimen, which incorporates a rich mixture of math, code, and science data. This focused training aims to enhance the model's reasoning abilities across these critical domains.

Key Capabilities

  • Enhanced Reasoning: Specifically designed to improve reasoning performance, as detailed in the paper "Think-at-Hard: Selective Latent Iterations to Improve Reasoning Language Models" (2511.08577).
  • Multi-domain Proficiency: Trained on a diverse dataset encompassing mathematics, programming code, and scientific texts, making it versatile for tasks within these areas.
  • Extended Context Window: Features a substantial context length of 32768 tokens, allowing it to process and understand longer inputs and more complex problem descriptions.

Good For

  • Mathematical Problem Solving: Its training includes significant math data, suggesting strong performance in numerical and logical reasoning.
  • Code Generation and Understanding: Exposure to code data makes it suitable for programming-related tasks.
  • Scientific Inquiry: Capable of assisting with tasks requiring understanding and generation of scientific content.
  • Research and Development: Ideal for applications where robust reasoning across technical and academic fields is crucial.