luganoquant/Nex-N2.5-mini

TEXT GENERATIONPricing:Input $0.4 / Cached $0.07 / Output $4Concurrent Unit Cost:2Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 16, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

Nex-N2.5-mini by Nex-AGI is a 35.1 billion parameter multimodal agentic model, part of the Nex-N2.5 family, designed for long-horizon tasks in real-world environments. It features enhanced computer use, web browsing, and visually grounded agentic capabilities, allowing it to act continuously and self-correct through visual feedback. This model excels at autonomously executing and testing programs, making it suitable for complex productivity tasks and scientific research.

Loading preview...

Nex-N2.5-mini: An Agentic Multimodal Model

Nex-N2.5-mini, developed by Nex-AGI, is a 35.1 billion parameter model from the Nex-N2.5 family, specifically engineered for long-horizon tasks in real-world environments. It builds upon the multimodal foundations of its predecessor, Nex-N2, with significant improvements in agentic capabilities.

Key Capabilities

  • Agentic Intelligence: Designed to act continuously and self-correct using visual feedback, enabling autonomous execution and testing of programs.
  • Multimodal Understanding: Integrates vision as a critical interface for perceiving environments and verifying outcomes.
  • Computer and Browser Use: Enhanced abilities for operating computers and navigating web browsers.
  • Task Coverage: Expanded range of agent training environments and task types, leading to gains in scientific research, knowledge work, and complex productivity tasks.

Performance Highlights

Nex-N2.5-mini demonstrates competitive performance across various benchmarks, particularly in agentic and multimodal tasks. While it is the 'mini' version, it still shows strong capabilities in areas like BrowseComp (83.4), OSWorld-Verified (71.2), and Vision2Web (52.9), indicating its proficiency in visually grounded interactions and web-based automation.

When to Use Nex-N2.5-mini

This model is particularly well-suited for applications requiring:

  • Automated Workflows: Tasks that involve interacting with computer interfaces or web browsers.
  • Agentic Systems: Developing agents that can perform multi-step tasks and adapt to real-world feedback.
  • Complex Problem Solving: Scenarios demanding continuous action and self-correction based on visual and environmental cues.
  • Research and Productivity: Aiding in scientific research and complex knowledge work where autonomous task execution is beneficial.

Finetunes
1 models