AnalysisAI Models
15 days ago
Coding benchmark: Step 3.7 and Qwen models compared
A simple coding benchmark tests Step 3.7 IQ4_XS, Qwen 3.5 122B-A10B, Qwen 3.6 27B, and Qwen 3.6 35B-A3B on a Go harness task. The author provides a detailed plan from Step 3.7 and compares model outputs.