abacusai/Dracarys2-72B-Instruct

Dracarys2-72B-Instruct is a 72.7 billion parameter instruction-tuned language model developed by Abacus.AI, finetuned from Qwen2.5-72B-Instruct. This model is specifically optimized for coding performance, demonstrating improved scores on the LiveCodeBench benchmark compared to its base model. With a context length of 131072 tokens, it excels in code generation and execution tasks.