Hanuman2/mergekit-slerp-xwepacy

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Jun 23, 2026Architecture:Transformer Featherless Exclusive Cold

Hanuman2/mergekit-slerp-xwepacy is a 7 billion parameter language model created by merging NousResearch/Hermes-2-Pro-Mistral-7B and WizardLM/WizardMath-7B-V1.1 using the SLERP method. This merge aims to combine the general instruction-following capabilities of Hermes-2-Pro with the mathematical reasoning strengths of WizardMath. It is designed for tasks requiring both robust conversational abilities and accurate numerical problem-solving within a 4096 token context length.

Loading preview...

Model Overview

This model, Hanuman2/mergekit-slerp-xwepacy, is a 7 billion parameter language model created through a merge operation using mergekit. It combines two distinct base models to leverage their respective strengths.

Merge Details

The model was constructed using the SLERP (Spherical Linear Interpolation) merge method. This technique blends the weights of the constituent models to create a new model with characteristics from both. The specific configuration used a V-shaped curve for the t parameter, emphasizing Hermes-2-Pro-Mistral-7B for input and output layers, and WizardMath-7B-V1.1 for the middle layers.

Constituent Models

  • NousResearch/Hermes-2-Pro-Mistral-7B: Known for its strong general instruction-following and conversational capabilities.
  • WizardLM/WizardMath-7B-V1.1: Specialized in mathematical reasoning and problem-solving.

Key Capabilities

By merging these models, Hanuman2/mergekit-slerp-xwepacy is intended to offer a balanced performance, combining:

  • General Instruction Following: Inherited from Hermes-2-Pro, enabling robust responses to a wide range of prompts.
  • Mathematical Reasoning: Enhanced by WizardMath, making it suitable for tasks involving numerical calculations and logical problem-solving.

Use Cases

This model is particularly well-suited for applications that require a blend of general conversational intelligence and specific mathematical aptitude, such as educational tools, technical support, or data analysis assistance where both understanding natural language and performing calculations are crucial.

Finetunes
2 models