[DHH Cross-Testing Programming Models: DeepSeek $23 Run, Fable Around $550] According to monitoring by Beating, Ruby on Rails creator DHH tested multiple models using the same Python-to-Rust task. DeepSeek Pro V4 Max ultimately completed the task in approximately 2.5 hours, with a total token cost of $23. Grok 4.6 took $55 and about 1.5 hours, while GPT-5.6 Sol cost $43. DeepSeek V4 Flash and GPT-5.6 Luna were unable to complete the task. The initial round was led by Fable, with DHH calculating costs at around $550 and a runtime of approximately 45 minutes. The task involved rewriting TerminalTextEffects from Python to Rust, using about 11 million tokens in total. Startup time was reduced from 87ms to 2ms, and rendering speed improved by 9.6 times. However, there is a caveat to the comparison: Fable first drafted a 483-line implementation plan, which was reviewed by Codex xhigh. The other models then directly followed this plan. Therefore, this is more of a comparison of the cost and efficiency of different models executing the same plan. The project has been open-sourced. The repository includes 354 consistency tests, which compare the Rust version and the original Python version frame by frame and byte by byte. The final output can be compiled into a static single file of approximately 3.3MB. [Original Link]