nvidia/Nemotron-Terminal-8B

Nemotron-Terminal-8B: Specialized for Autonomous Terminal Interaction

Nemotron-Terminal-8B is an 8 billion parameter model developed by NVIDIA, fine-tuned from the Qwen3 architecture. It is part of the Nemotron-Terminal family, which is specifically designed for autonomous terminal interaction.

Key Capabilities & Features

• Specialized Fine-tuning: Trained using the Nemotron-Terminal-Corpus, a large-scale open-source dataset focused on terminal tasks.
• Enhanced Performance: Demonstrates significant improvements over its Qwen3 baseline on the Terminal-Bench 2.0 benchmark, achieving 13.0% accuracy compared to the base model's 2.47%.
• Structured Output: Designed to output structured JSON, including analysis, plan, and commands for executing terminal operations.
• Terminus 2 Scaffolding: Utilizes the Terminus 2 scaffolding for training and is recommended for consistent evaluation.

Use Cases

This model is ideal for applications requiring automated command-line execution, intelligent terminal agents, and systems that need to analyze and interact with terminal environments programmatically. Its specialization allows it to rival the performance of much larger, more general-purpose models in its domain.