13 comments

  • underyx5 hours ago
    &gt; the slow performance decays<p>the decays are just more capable other models entering the population, making all prior models lose more frequently
    • TekMol2 hours ago
      No, that is not how ELO scores work.
      • qnleigh2 hours ago
        As far as I understand, this is exactly how ELO scores work. If a more capable show up and starts beating all the other models, it literally takes ELO points from everyone else.<p><a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Elo_rating_system" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Elo_rating_system</a>
        • TekMol6 minutes ago
          <p><pre><code> If a more capable show up and starts beating all the other models </code></pre> There is an instance of this in the chart. In 2025-06-24 when Gemini-2.5-pro shows up. As you can see, the ELO of the others do not drop.
        • harperlee1 hour ago
          Depends on the test design; is an agent competing against other agent in a given match, or against a test? Plus! Does the test&#x27;s ELO fluctuate?
      • tasuki49 minutes ago
        Yes, that is in fact how Elo can work[0]. There are quite many ways Elo systems can work.<p>[0]: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Elo_rating_system" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Elo_rating_system</a>
      • whiplash4512 hours ago
        It depends what you use as an anchor. If the anchor is a fixed model, you’re right. If the anchor is updated to a better model over time, then the elo of historical models degrades, right?
  • tedsanders5 hours ago
    For what it&#x27;s worth, I work at OpenAI and I can guarantee you that we don&#x27;t switch to heavily quantized models or otherwise nerf them when we&#x27;re under high load. It&#x27;s true that the product experience can change over time - we&#x27;re frequently tweaking ChatGPT &amp; Codex with the intention of making them better - but we don&#x27;t pull any nefarious time-of-day shenanigans or similar. You should get what you pay for.
    • selcuka5 hours ago
      &gt; we don&#x27;t switch to heavily quantized models<p>That sounded like a press bulletin, so just to let you clarify yourself: Does that mean you may switch to <i>lightly</i> quantized models?
      • jychang4 hours ago
        There&#x27;s almost 0% chance that OpenAI doesn&#x27;t quantize the model right off the bat.<p>I am willing to bet large amounts of money that OpenAI would never release a model served as fully BF16 in the year of our lord 2026. That would be insane operationally. They&#x27;re almost certainly doing QAT to FP4 for FFN, and a similar or slightly larger quant for attention tensors.
        • selcuka4 hours ago
          It&#x27;s ok if they never release a BF16 model, but it&#x27;s less ok if they release it, win the benchmarks, then quantise it after a few weeks.
          • retinaros38 minutes ago
            that is for sure what everyone does. also they train on evals with the datasets that they would be bench against.
    • Ciph4 hours ago
      Thank you for your answer. I have a similar question as OP, but in regards of the GPT models in MS copilot. My experience is that the response quality is much better when calling the API directly or through the webUI.<p>I know this might be a question that&#x27;s impossible for you to answer, but if you can shed any light to this matter, I&#x27;d be grateful as I am doing an analysis over what AI solutions that can be suitable for my organisation.
      • sans_souse2 hours ago
        As phrased the only answer is the question; &quot;as opposed to what?&quot;
      • aiscoming2 hours ago
        webUIs have giant system prompts built in<p>APIs have much smaller ones
    • _kidlike2 hours ago
      its very interesting to see that this only happens to American companies. What gives?
  • ponyous2 hours ago
    Seems like Chinese labs are the only ones that are trustworthy (at least when it gets to this specific issue). This feels so ironic haha
    • mordae2 hours ago
      I am using novita-hosted DeepSeek V4 (Flash) for work and DeepSeek API for personal projects.<p>Novita&#x27;s has occassional problem counting white space. DeepSeek hosted does not.<p>No idea why.
    • lukewarm70738 minutes ago
      there is something greatly trustworthy about open source
  • whiplash4512 hours ago
    Neat. Would you add the option to normalize the elo over time (e.g update the model used as an anchor for the elo computation) so the diff between labs is more visible?
  • eis5 hours ago
    The Elo rating system measures <i>relative</i> performance to the other models. As the other models improve or rather newer better models enter the list, the Elo score of a given existing model will tend to decrease even though there might be no changes whatsoever to the model or its system prompt.<p>You can&#x27;t use Elo scores to measure decay of a models performance in absolute terms. For that you need a fixed harness running over a fixed set of tests.
    • bob10292 hours ago
      The relative and auto-scaling nature of Elo ranking feels like an advantage here.<p>Relative ranking systems extract more information per tournament. You will get something approximating the actual latent skill level with enough of them.
      • eis2 hours ago
        Advantage for what exactly though? I&#x27;m not saying Elo Ranking doesn&#x27;t give any information. It just doesn&#x27;t give the information that the OP&#x27;s project claims to be able to give: that models get nerfed over time. You could extract this kind of information from the raw results of each evaluation round between two models, ignoring any new model entries and compare these over time but not from the resulting Elo scores with an ever changing list of models.<p>New models are on average better than older models, the average skill of the population of models increases over time and so you are mathematically guaranteed that any existing model will over time degrade in Elo score even though it didn&#x27;t change itself in any way.<p>It&#x27;s like benchmarking a model against a list of challenges that over time are made more and more difficult and then claiming the model got nerfed because its score declined.<p>Elo is good at establishing an overall ranking order across models but that&#x27;s not what this is about.<p>To detect nerfing of a model, projects like <a href="https:&#x2F;&#x2F;marginlab.ai&#x2F;trackers&#x2F;claude-code&#x2F;" rel="nofollow">https:&#x2F;&#x2F;marginlab.ai&#x2F;trackers&#x2F;claude-code&#x2F;</a> are much much better (I&#x27;m not affiliated in any way).
    • TurdF3rguson1 hour ago
      Is that strictly true? ELO rankings do also inflate over time (looking at you, Chess GMs)
      • tasuki43 minutes ago
        Elo systems often include one or more ways new points can enter the system. The system used by the European Go Federation has three ways iirc: 1. Cannot go under 100, 2. Cannot lose more than 100 points in one tournament, 3. Weaker player beating a stronger one (which is countered by the stronger player beating the weaker one, but it&#x27;s not balanced: if two people only play each other forever and ever, both of their Elos will grow).
  • cherioo4 hours ago
    The interesting thing I find is how Anthropic has been more consistently improving over time in the last few years, that allows it to catchup and surpass OpenAI and Google. The latter two have pretty much plateau over the last year or so. GPT 5.5 is somehow not moving the needle at all.<p>I hope to see the other labs can bring back competition soon!
    • XCSme3 hours ago
      Gpt 5.5 is quite a big leap, it&#x27;s a lot better than opus 4.7 for agentic coding
      • energy1233 hours ago
        Arena only allows very small context sizes, so it&#x27;s a noisy benchmark for what we care about IRL.
      • mettamage2 hours ago
        Better in what ways? I&#x27;m just curious about your experience.
        • XCSme2 hours ago
          Consistency, not making mistakes.
          • mettamage2 hours ago
            Ahh... that is indeed an issue I have with Claude. I&#x27;ll check it out!
  • kimjune013 hours ago
    Although Arena is adversarial and resistant to goodharting, it&#x27;s not immune. Models that train on Arena converge on helpfulness, not necessarily truthiness
  • tedsanders5 hours ago
    FYI, Elo isn&#x27;t an acronym - it&#x27;s a person&#x27;s name. No need to capitalize it as ELO.
    • alex_duf2 hours ago
      Electric Light Orchestra anyone?
    • andrewshadura3 hours ago
      Unless you&#x27;ve just missed your last train to London.
    • SilverElfin5 hours ago
      You’re right: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Elo_rating_system" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Elo_rating_system</a>
      • gitowiec4 hours ago
        &gt; ELO ratings<p>Thank you, I just looked at the chart and said to myself: ELO? YOLO!<p>That Elo ranking is also called chess ranking
        • andrewshadura3 hours ago
          Élő. Meaning alive (él = it lives, -ő = adjective)
  • jdw643 hours ago
    This is great, but personally, I really wish we had an Elo leaderboard specifically for the quality of coding agents.<p>Honestly, in my opinion, GPT-5.5 Codex doesn&#x27;t just crush Claude Code 4.7 opus —it&#x27;s writing code at a level so advanced that I sometimes struggle to even fully comprehend it. Even when navigating fairly massive codebases spanning four different languages and regions (US, China, Korea, and Japan), Codex&#x27;s performance is simply overwhelming.<p>How would we even go about properly measuring and benchmarking the Elo for autonomous agents like this?
    • vachanmn1233 hours ago
      Isn&#x27;t code that you fail to understand literally a sign that its worse?
      • jdw643 hours ago
        It was often much faster, and when I revisited the code later, there were cases where I realized it had moved the implementation toward a better abstraction.
      • jdw643 hours ago
        I should also add that I am not claiming to be a particularly great programmer. I have never worked at FAANG, and I haven&#x27;t had much exposure to the kind of massive codebases those engineers deal with every day.<p>Most of the code I&#x27;ve worked with comes from Korean and Chinese startups, industrial contractors, or older corporate research-lab environments. So I know my frame of reference is limited.<p>When I write code, I usually rely on fairly conservative patterns: Result-style error handling instead of throwing exceptions through business logic, aggressive use of guard clauses, small policy&#x2F;strategy objects, and adapters at I&#x2F;O boundaries. I also prefer placing a normalization layer before analysis and building pure transformation pipelines wherever possible.<p>So when Codex produced a design that decoupled the messy input adapter from the stable normalized data, and then separated that from the analyzer, it wasn&#x27;t just &#x27;fancier code.&#x27; It aligned perfectly with the architectural direction I already value, but it pushed the boundaries of that design further than I would have initially done myself.<p>This is exactly why I hesitate to dismiss code as &#x27;bad&#x27; just because I don&#x27;t immediately understand it. Sometimes, it really is just bad code. But sometimes, the abstraction is simply a bit ahead of my current local mental model, and I only grasp its true value after a second or third requirement is introduced.<p>To be completely honest, using AI has caused a significant drop in my programming confidence. Since AI is ultimately trained on codebases written by top-tier programmers, its output essentially represents the average of those top developers—or perhaps slightly below their absolute peak.<p>I often find myself realizing that the code I write by hand simply cannot beat it
  • fph2 hours ago
    Very neat! It would be great to extend it to non-flagship models as well.
  • Thomashuet4 hours ago
    It seems to be a USA only thing, Chinese models and Mistral don&#x27;t show any downward trend.
    • TurdF3rguson25 minutes ago
      Sure they do. Most models are on a downward trend because newer models are moving into top spots.
    • patall3 hours ago
      Wouldn&#x27;t it be really weird if a open-weight model dropped in performance? Because then, it would rather be the Elo ranking
  • refulgentis4 hours ago
    Is this slop? It has wildly aggressive language that agrees with a subset of pop sentiment, re: models being “nerfed”. It promises to reveal this nerfing. Then, it goes on to…provide an innocuous mapping of LM Arena scores that always go up?
    • ninjalanternshk1 hour ago
      It links to the GitHub repo for the project, and while it’s not inconceivable that an AI bot would create and populate a functioning public GitHub repo, it’s pretty unlikely.
  • gptbased1 hour ago
    [dead]