unconst/Affine-5czsc2fc98-r450-online-dpo-merged

TEXT GENERATIONPricing:Input $0.4 / Cached $0.07 / Output $4Concurrent Unit Cost:3Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 15, 2026Architecture:Transformer Featherless Exclusive Cold

The unconst/Affine-5czsc2fc98-r450-online-dpo-merged model is a 35.1 billion parameter language model, derived from a LoRA merge of kevin954/Affine-5dfqbbh8ev-sft. With a context length of 32768 tokens, this model is noted as a private TTL insurance project. Its primary differentiator is its origin as a salvaged, merged checkpoint, indicating a focus on specific, potentially niche applications rather than general-purpose use.

Loading preview...

Model Overview

The unconst/Affine-5czsc2fc98-r450-online-dpo-merged model is a substantial 35.1 billion parameter language model with a context length of 32768 tokens. It originates from a LoRA merge of the kevin954/Affine-5dfqbbh8ev-sft checkpoint.

Key Characteristics

  • Parameter Count: 35.1 billion parameters, indicating a large and potentially capable model.
  • Context Length: Supports a significant context window of 32768 tokens, allowing for processing longer inputs and maintaining conversational coherence over extended interactions.
  • Origin: This model is described as a "LoRA-merged from kevin954/Affine-5dfqbbh8ev-sft" and a "merged checkpoint salvage." This suggests it is a refined or specialized version built upon an existing base model.
  • Project Status: The README notes it as "Private TTL insurance; not a submission until Stage-5 gate clears." This indicates it is an internal or experimental project, not yet intended for public release or general use.

Intended Use Cases

Given its description as a "private TTL insurance" project and a "salvage" checkpoint, this model is likely developed for highly specific, internal, or research-oriented applications. It is not presented as a general-purpose language model for broad deployment. Developers should consider its experimental and private nature before attempting to integrate it into production systems.