shrugging-shoulders/Amberlight-Lux-12B

TEXT GENERATIONPricing:Input $0.87 / Cached $0.2 / Output $0.99Concurrent Unit Cost:1Model Size:12BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 22, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Amberlight-Lux-12B is a 12 billion parameter language model developed by shrugging-shoulders, building upon the Amberlight-12B architecture. This iteration is further trained to enhance stability and reduce error proneness, making it a more reliable option for general text generation tasks. It supports a context length of 32768 tokens and is configured for ChatML template format, suitable for conversational applications.

Loading preview...

Amberlight-Lux-12B Overview

Amberlight-Lux-12B is an enhanced version of the original shrugging-shoulders/Amberlight-12B model. This 12 billion parameter model has undergone additional training to improve its overall stability and reduce the likelihood of generating errors, offering a more robust performance for various language tasks.

Key Characteristics

  • Enhanced Stability: Specifically trained to be less prone to errors and exhibit greater stability compared to its predecessor.
  • Context Length: Supports a substantial context window of 32768 tokens, allowing for processing longer inputs and generating more coherent, extended outputs.
  • ChatML Format: Utilizes the ChatML template format, making it well-suited for integration into chat-based applications and conversational AI systems.

Recommended Settings

While flexible, the model has been tested with the following default generation parameters:

  • Temperature: 0.8
  • Repetition Penalty: 1.05
  • TOP_P: 0.90
  • MIN_P: 0.025

These settings are suggested for optimal performance, though temperatures between 0.6 and 1.0 have also shown reasonable results. The model benefits from the contributions of Team mradermacher for providing GGUF format quantizations.

Finetunes
1 models