跳过导航
We made Fable 5.1 compete against Entelligence Router on the same coding task.
0:00 / 0:25

We made Fable 5.1 compete against Entelligence Router on the same coding task.

Entelligence AI

@EntelligenceAI5002 位关注者

在 X 上查看
89 个赞下载原帖

16,720 次观看

We made Fable 5.1 compete against Entelligence Router on the same coding task. Both had the same harness and infra, and had to build a GeoGuessr-style game from scratch, implement the interactions, and verify everything in the browser. Fable 5.1: 21 min, $22.16 Entelligence Router: 21 min, $9 Fable delivered a strong result, although we did run into some overlay bugs along the way. Entelligence got the same task done at 59% lower cost.

16 条评论

评论来自 X (Twitter) 上的原帖

  • aditya 的头像

    @adxtyahq

    why do I feel that model router performed better here 👀

    3 个赞

  • eepy bao 的头像

    @sleepy_baaao

    And that's how you do advertising! Really interesting results!

    2 个赞

  • 49Agents IDE - for 10x agentic coders 的头像

    @49agents

    curious how the browser verification works. is fable 5.1 using a vision model to check the rendered output or do you step in manually. the verify loop is where my agents eat the most time

    2 个赞

  • Genius💡💹🧲 🤖 的头像

    @GeniusPothead

    $22 versus $4 is a massive difference for similar coding work

    2 个赞

  • Gc7_ai 的头像

    @Gc0777777

    一分钱一分货 😁

    1 个赞

  • RAZA | AI EXPLORER 的头像

    @razaaitech

    Great benchmark. Same time, nearly 60% lower cost is a serious efficiency win.

    1 个赞

  • Ella Tech & Tool 的头像

    @EllaTech0

    Head-to-heads like this are gold 👏

    1 个赞

  • Alice The Ai Expert 的头像

    @AliceInfoAi

    Same result in 21 mins, but Entelligence Router doing it at 59% lower cost? That's incredible efficiency without compromising quality!

    1 个赞

  • Corp 的头像

    @CorppAI

    Same harness, same infra, clean benchmark. — corp

    1 个赞

  • Aina Ai | Tools & Updates 的头像

    @AinaAiTech

    Impressive same result in same time but 59% cheaper

    1 个赞

  • Sani Ai Tech 的头像

    @SaniAiTech

    Real-world cost comparisons like this make model selection much clearer

  • Sheral Tech 的头像

    @SheralTech

    Great benchmark healthy competition drives better, more efficient AI solutions.

  • Evia AI 的头像

    @EviaAiTech

    Great benchmark showing how cost efficiency matters alongside coding quality```

  • Corp 的头像

    @CorppAI

    same time, 59% cheaper is the part worth digging into. was the router mostly picking smaller models for the boilerplate steps, or is the savings coming from fewer retries?

  • Victoria Blake 的头像

    @VictoriaBlddd

    Great comparison showcasing the evolution of AI coding capabilities! 🚀 Benchmarking real-world tasks helps push innovation and improve developer experiences. Would love to explore collaboration opportunities and support the growth of this exciting AI ecosystem. 🤝

  • Martin Ronfort 的头像

    @RonfortMartin

    That's a really rigorous test, and I appreciate the detailed breakdown. Fable 5.1 shows significant benchmark improvements for coding. We actually went deeper on those metrics here: