sargevinix/archai-v1-merged

TEXT GENERATIONConcurrency Cost:1Model Size:12BQuant:FP8Ctx Length:32kPublished:May 8, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

sargevinix/archai-v1-merged is a 12 billion parameter Mistral-based causal language model developed by sargevinix, fine-tuned from unsloth/mistral-nemo-instruct-2407-bnb-4bit. This model was trained using Unsloth and Huggingface's TRL library, enabling 2x faster fine-tuning. With a 32768 token context length, it is optimized for efficient performance on tasks typically handled by Mistral-family models.

Loading preview...

Model Overview

sargevinix/archai-v1-merged is a 12 billion parameter language model, fine-tuned by sargevinix. It is based on the Mistral architecture, specifically fine-tuned from the unsloth/mistral-nemo-instruct-2407-bnb-4bit model. A key characteristic of this model's development is its training methodology, which leveraged Unsloth and Huggingface's TRL library to achieve a 2x speedup in the fine-tuning process.

Key Characteristics

  • Architecture: Mistral-based, fine-tuned from unsloth/mistral-nemo-instruct-2407-bnb-4bit.
  • Parameter Count: 12 billion parameters.
  • Context Length: Supports a substantial context window of 32768 tokens.
  • Training Efficiency: Fine-tuned with Unsloth, resulting in significantly faster training times.
  • License: Released under the Apache-2.0 license.

Potential Use Cases

This model is suitable for applications requiring a Mistral-family model with a large context window, where efficient fine-tuning was a priority during development. Its foundation suggests capabilities for general language understanding and generation tasks, benefiting from the Mistral architecture's known performance characteristics.