Aznaur/tbench-qwen-sft-multitask-nat-v11
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:Jan 24, 2026Architecture:Transformer Cold

Aznaur/tbench-qwen-sft-multitask-nat-v11 is an 8 billion parameter Qwen3-8B model fine-tuned by Aznaur with a 32768-token context length. It utilizes Negative-Aware Training (NAT) v11, specifically designed to understand and avoid common failure patterns in terminal command execution. This model excels at generating robust and error-resistant terminal commands by learning from both successful and unsuccessful examples.

Loading preview...