Ternary Bonsai 27B

Ternary Bonsai 27B is a large language model architecture evaluated for performance against Qwen 27B variants. Recent benchmarking highlights its behavior under different quantization and draft configurations.

Performance Benchmarking

Recent analysis compares Ternary Bonsai 27B against the Qwen 3.6 27B MTP model in a local 16GB setup. Key findings include:

References