DH Transverse Programming Model: DeepSeek 23 ran, Fable approximately $550

DHH (author of Ruby on Railways) conducted a programming model test, which compared multiple models using the Python-Rust mission. DeepSeek Pro V4 Max was finalized for approximately 2.5 hours at a total cost of $23. Grok 4.6 costs $55, or approximately 1.5 hours, while GPT-5.6 SOL costs $43. DeepSeek V4 Flash and GPT-5.6 Luna were unable to complete their tasks. The first round was led by Fable, with a statistical cost of about $550 per DH, which took about 45 minutes. The project has been open-sourced and the warehouse contains 354 consistency tests, frame-by-story comparison of Rust and Python versions, and final products can be compiled into some 3.3 MB static single files。