CloudGoat/Mephisto-1.0-4B

VISIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:4.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 2, 2026Architecture:Transformer0.0K Featherless Exclusive Cold

CloudGoat/Mephisto-1.0-4B is a 4.5 billion parameter agentic language model developed by CloudGoat, built upon the Qwen3.5-4B architecture. This model was created using a multi-stage merging process with a forked version of Mergekit, combining several fine-tuned models including Jackrong/Qwopus3.5-4B-v3, Jackrong/Qwopus3.5-4B-Coder, BAAI/AREX-Turbo, and InternScience/Agents-A1-4B. Its primary differentiator is its sophisticated merging methodology, designed to enhance agentic capabilities, making it suitable for tasks requiring complex reasoning and multi-step interactions.

Loading preview...

Overview

CloudGoat/Mephisto-1.0-4B is a 4.5 billion parameter agentic language model based on the Qwen3.5-4B architecture. Developed by CloudGoat, this model distinguishes itself through a sophisticated, multi-stage merging process using a specialized version of Mergekit. Unlike single-step merges, Mephisto-1.0-4B integrates capabilities from multiple fine-tuned models, including those focused on coding and agentic behaviors, to create a more robust and capable agent.

Key Capabilities

  • Enhanced Agentic Performance: Designed to excel in tasks requiring agentic reasoning and multi-step problem-solving due to its unique merging strategy.
  • Multi-Model Integration: Combines strengths from various specialized models like Jackrong/Qwopus3.5-4B-v3, Jackrong/Qwopus3.5-4B-Coder, BAAI/AREX-Turbo, and InternScience/Agents-A1-4B.
  • Advanced Merging Methodology: Utilizes a multi-stage SLERP and DARE-TIES merging approach, which is more complex than typical single-step merges, aiming for superior performance integration.

Good for

  • Agent-based Applications: Ideal for developing AI agents that require advanced reasoning and interaction capabilities.
  • Complex Task Automation: Suitable for scenarios where a model needs to perform multi-step operations or integrate information from various domains.
  • Research into Model Merging: Provides a practical example of advanced merging techniques for researchers and developers interested in creating composite models.

Finetunes
1 models