allenai/qwen35-9b-termigen

VISIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 19, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The allenai/qwen35-9b-termigen model is a 9 billion parameter language model developed by Ai2, fine-tuned from Qwen 3.5 9B. It is specifically optimized as a terminal-agent, trained using DPPO on the TermiGen dataset. This model excels at interacting with and executing commands within a terminal environment, making it suitable for automated system administration and command-line task execution. It features a 32768 token context length, designed for complex terminal interactions.

Loading preview...

Model Overview

allenai/qwen35-9b-termigen is a 9 billion parameter language model developed by Ai2, specifically fine-tuned from Qwen 3.5 9B to function as a terminal-agent. This model is part of a collection of terminal agents and was trained using DPPO on the TermiGen dataset.

Key Capabilities

  • Terminal Agent Functionality: Designed to interact with and execute commands within a terminal environment.
  • DPPO Fine-tuning: Utilizes Deep Proximal Policy Optimization for enhanced performance in terminal-based tasks.
  • Context Length: Supports a maximum overall context of 65536 tokens, with max per-turn tokens of 16384, suitable for managing complex command sequences.
  • Evaluation Performance: Achieved a score of 49.4 ± 1.5 on TB Lite and 25.1 ± 1.9 on TB 2.1, demonstrating its proficiency in terminal-based benchmarks.

Use Cases

  • Automated System Administration: Ideal for automating routine command-line tasks and system management.
  • Command-Line Task Execution: Can be used to interpret and execute sequences of terminal commands.
  • Research in Agentic AI: Serves as an ablation model for studying terminal agents, as detailed in the Tmax paper.

This model is licensed under Apache 2.0 and is intended for research and educational use.