unconst/Affine-5czsc2fc98-r515-offline-dpo-hialpha-hirank-lobeta-softctx-ultraextrasteps-merged

TEXT GENERATIONConcurrent Unit Cost:3Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 16, 2026Architecture:Transformer Featherless Exclusive Cold

The unconst/Affine-5czsc2fc98-r515-offline-dpo-hialpha-hirank-lobeta-softctx-ultraextrasteps-merged model is a 35.1 billion parameter language model, developed by unconst, that has been LoRA-merged from kevin954/Affine-5dfqbbh8ev-sft. This model is an internal checkpoint, specifically a private TTL insurance, and is not intended as a public submission until its Stage-5 gate clears. It features a context length of 32768 tokens, indicating its capability for processing extensive inputs.

Loading preview...

Model Overview

The unconst/Affine-5czsc2fc98-r515-offline-dpo-hialpha-hirank-lobeta-softctx-ultraextrasteps-merged is a substantial 35.1 billion parameter language model. It was developed by unconst and represents a LoRA-merged checkpoint, originating from kevin954/Affine-5dfqbbh8ev-sft. This model is characterized by its large context window of 32768 tokens, allowing it to handle extensive textual inputs and maintain coherence over long conversations or documents.

Key Characteristics

  • Parameter Count: 35.1 billion parameters, indicating a high capacity for complex language understanding and generation tasks.
  • Context Length: Supports a significant context window of 32768 tokens, beneficial for applications requiring deep contextual awareness.
  • Development Status: This is an internal, private checkpoint, described as "TTL insurance," and is not yet a public submission. It is awaiting clearance of its Stage-5 gate.

Intended Use Cases

Given its internal and private nature, this model is currently intended for specific, internal development and evaluation purposes by its creators. It is not yet released for general public use or external applications. Developers should monitor its status for potential future public availability and specific use case recommendations once its development gates are cleared.