10 comments

  • thorum2 minutes ago
    Interesting that all four models converge on such similar designs, for such short prompts.
  • singingtoday15 minutes ago
    Love the idea, I think more complex games would show the gap in ability better.<p>Do it again but this time get them to make a multiplayer online Jetmen REVIVAL game. Online play is key, because it&#x27;s very complex. Jetmen is a good game for this since it has physics and customization that&#x27;s complex enough but still simple.
  • dadoum24 minutes ago
    I tried to one-shot the first test (the Rubik&#x27;s Cube test) with LucidQuery&#x27;s Swift model, to test it, as there are not much benchmarks about it and that they brag a lot about it, and I was pleasantly surprised to see it achieving a result similar to Grok 4.5 but in one shot (there is the same issue that if you scramble twice the solve button does not work anymore, but it got it in one shot).<p>Though it crunched most of the free quota, 47111 tokens, so I couldn&#x27;t make multiple attempts.
  • Kuyawa4 minutes ago
    I&#x27;d like to see the comparisons with DeepSeek, Qwen, Mimo, Kimmi and GLM
  • mlmonkey41 minutes ago
    I am 99% sure the post was written by AI
    • jszymborski30 minutes ago
      The honest takeaway: this is 100% written by an LLM.
      • mlmonkey8 minutes ago
        That was the honest giveaway ... :-D
    • dwa359233 minutes ago
      i will give you the remaining 1% because i felt the same way.
  • krauq11 minutes ago
    Too nice to Grok, if there are really cost savings it should say how much each of the three demos cost so we can judge if it&#x27;s worth the lower quality (probably not). The time to complete each would also be interesting.
  • jeffgreco25 minutes ago
    So strange to write a whole post with Claude giving the best results and Grok consistently the worst, but awarding Grok the winner because at least it did the worst fastest?
    • singingtoday18 minutes ago
      GPT was the worst on the Rubik&#x27;s cube
      • GaggiX9 minutes ago
        Grok did not render anything, they had to prompt it again.
  • paxys44 minutes ago
    Why not wait <i>one</i> more day for GPT-5.6?
    • foxfired19 minutes ago
      If we wait for the next models, we will never test anything because there will always be another model. Like the Ai Scotsman:<p>&gt; &quot;Nay, laddie, that’s no’ the real AI Scotsman! He’s grander still! More powerful! Just wait for the next model!&quot;
    • acters18 minutes ago
      I worry that GPT 5.6 will be heavily restricted and have the same feature to fallback to another model like Claude fable 5 does all too often. That fallback shenanigans mess up actual benchmarks and I don&#x27;t like it.
    • m4rkuskk41 minutes ago
      That will be in the Part 2 article.
    • faitswulff27 minutes ago
      Also throw in GLM 5.2 for good measure
    • fluidcruft30 minutes ago
      And why not Sonnet?
  • maxdo54 minutes ago
    Tried at work , this release def a moment I will remember. My work is not the same . The model is the first model that offer exactly as I want :<p>For hard tasks , that needs precision I will wait and pay expensive tokens<p>For everything else , query data , logs, rolling out releases , I’m using grok and it’s much better vs other tools and much cheaper too .
  • RickS34 minutes ago
    Barring the retry thing, n=1 on all models? Am I misreading, or is this a joke?<p>Variance in quality on these things is so, so high.