Macvex/albedo-unc

TEXT GENERATIONPricing:Input $0.4 / Cached $0.07 / Output $4Concurrent Unit Cost:3Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 16, 2026Architecture:Transformer Featherless Exclusive Cold

Macvex/albedo-unc is a 35.1 billion parameter language model, merged from kevin954/Affine-5dfqbbh8ev-sft. This model is a private TTL insurance checkpoint, not yet a public submission, and is intended for specific, internal use cases. Its primary characteristic is its origin as a LoRA-merged checkpoint, indicating a fine-tuned or adapted version of a base model.

Loading preview...

Overview

Macvex/albedo-unc is a 35.1 billion parameter language model that originated as a LoRA-merged checkpoint from kevin954/Affine-5dfqbbh8ev-sft. This model is currently designated as a private TTL insurance asset, meaning it is not intended for general public use or as a final submission until it clears a Stage-5 gate.

Key Characteristics

  • Parameter Count: 35.1 billion parameters, indicating a substantial model size capable of complex language understanding and generation tasks.
  • Origin: Developed through a LoRA (Low-Rank Adaptation) merge, suggesting it has been fine-tuned or adapted from a base model for specific performance characteristics.
  • Status: Marked as a private TTL insurance checkpoint, implying it serves as a backup or intermediate version within a development pipeline.

Use Case Considerations

Given its status as a private, LoRA-merged checkpoint, Macvex/albedo-unc is likely intended for:

  • Internal Development: For testing, iteration, or specific project requirements within the developer's scope.
  • Specialized Tasks: The LoRA merge suggests optimization for particular domains or tasks, though specifics are not detailed in the README.

Users should note that this model is not presented as a finalized or publicly released version and its capabilities are tied to its specific development context.