DHH Tests Programming Model Costs and Efficiency: DeepSeek at $23, Fable at $550

DHH tested multiple models for the Python to Rust task. DeepSeek Pro V4 Max took about 2.5 hours, with a total token cost of $23. Grok 4.6 cost $55 and took 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 of statistics led by Fable had a cost of approximately $550 and took about 45 minutes, rewriting TerminalTextEffects from Python to Rust, using around 11 million tokens, reducing startup time from 87ms to 2ms, and improving rendering speed by 9.6 times. The comparison is based on Fable first writing 483 lines of implementation, which were reviewed by Codex xhigh, allowing other models to directly use this plan to compare the costs and efficiencies of executing the same plan across different models. The project has been open-sourced, and the repository includes 354 consistency tests, comparing frame-by-frame and byte-by-byte between the Rust version and the original Python version, with the final product being compilable into a static single file of about 3.3MB.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.