groxaxo/MagiSeek-Pro-V1

TEXT GENERATIONPricing:Input $0.7 / Cached $0.04 / Output $1.16Concurrent Unit Cost:2Model Size:24BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 7, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

groxaxo/MagiSeek-Pro-V1 is a 24 billion parameter Mistral-architecture model checkpoint, published by groxaxo, with a native context length of up to 131,072 tokens. It is derived from WarlordHermes/Magidonia-24B-v4.3-creative-ORPO and fine-tuned through a 5-phase QLoRA curriculum, focusing on creative writing, instruction following, and agentic/tool-use tasks. This bf16 merged full-precision model excels at long-context reasoning and is suitable as a source for further quantization.

Loading preview...

Overview of MagiSeek-Pro-V1

MagiSeek-Pro-V1 is a 24 billion parameter model based on the Mistral architecture, developed by groxaxo. It is a bf16 merged checkpoint, originating from WarlordHermes/Magidonia-24B-v4.3-creative-ORPO, and has undergone a unique 5-phase QLoRA training curriculum. This process has imbued the model with distinct capabilities, particularly in creative generation and advanced reasoning.

Key Capabilities

  • Extended Context Handling: Supports a native context length of up to 131,072 tokens, enabling deep understanding and generation for long-form content.
  • Agentic Reasoning: Features a specialized Phase 4 training using real tool-execution transcripts and Claude Opus 4.6 reasoning traces, enhancing its ability to call and utilize tools effectively.
  • DeepSeek-style Reasoning: Phase 5 training focused on sharpening step-by-step reasoning, allowing for more structured and logical outputs without compromising its creative foundation.
  • Full Precision: Released as a bf16 merged model, ensuring full precision and serving as an optimal base for subsequent quantization (e.g., GGUF, GPTQ, AWQ).
  • Instruction Following & Creative Writing: Combines the creative instincts of its Magidonia base with refined instruction-following abilities.

Intended Use Cases

  • General Instruction Following: Capable of understanding and executing a wide range of instructions.
  • Creative Writing: Excels in generating imaginative and coherent text.
  • Agentic/Tool-Use Tasks: Designed for scenarios requiring the model to interact with external tools and perform complex, multi-step reasoning.
  • Long-Context Applications: Ideal for tasks that benefit from processing and generating extensive textual information.

Finetunes
1 models