EldritchLabs/MN-Aura-12B-v1

TEXT GENERATIONPricing:Input $1.2 / Output $4.8Concurrent Unit Cost:1Model Size:12.2BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 25, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

EldritchLabs/MN-Aura-12B-v1 is a 12.2 billion parameter language model created by EldritchLabs, merged using the Adaptive Unified Riemannian Annealing (AURA) method. This model integrates several pre-trained models, including Mistral-Nemo-Base-2407-ChatML as its base, to combine their strengths. It is designed for general language tasks, leveraging its merged architecture for broad applicability.

Loading preview...

Overview

EldritchLabs/MN-Aura-12B-v1 is a 12.2 billion parameter language model developed by EldritchLabs. It was constructed using the novel Adaptive Unified Riemannian Annealing (AURA) merge method, which is part of the mergekit-exp framework. The model's foundation is Retreatcost/Mistral-Nemo-Base-2407-ChatML, and it incorporates contributions from five other distinct 12B models, including DreadPoor/Famino-12B-Model_Stock, EldritchLabs/MN-Starlight-Sylph-12B, OccultAI/MN-Nazgul-12B-v1, shrugging-shoulders/Amberlight-Lux-12B, and WokeAI/Tankie-DPE-12B-SFT-v2. This merging strategy aims to synthesize the capabilities of multiple specialized models into a single, more versatile entity.

Key Capabilities

  • Merged Architecture: Leverages the AURA merge method to combine the strengths of six different 12B parameter models.
  • Base Model: Built upon the Retreatcost/Mistral-Nemo-Base-2407-ChatML, providing a strong foundation for chat-oriented tasks.
  • Parameter Count: Operates with 12.2 billion parameters, offering a balance between performance and computational requirements.
  • Context Length: Supports a context window of 32768 tokens, enabling processing of longer inputs and maintaining conversational coherence over extended interactions.

Good For

  • General Language Tasks: Suitable for a wide range of applications due to its diverse merged components.
  • Exploration of Merged Models: Ideal for developers interested in experimenting with models created via advanced merging techniques like AURA.
  • Applications Requiring Moderate Scale: Provides a capable solution for use cases that benefit from a 12B parameter model with a substantial context window.

Finetunes
1 models