Researchers asked frontier models to play chess. Have a look at the MAR rates in Table 3. When not explicitly told which moves were legal, no model identified legal moves at a rate better than 80%. Many asked for more illegal moves than legal moves. And even when explicitly told which moves were legal, the models continued to ask for illegal moves. With illegal asks discarded, none of the bots could beat a chess model calibrated to 1100 ELO.
The author of the originating post says that "current frontier models need laborious oversight and guardrails on even the simplest tasks", and he's absolutely correct.
> Gemini 2.5 Pro, O3, Claude Sonnet 3.7 and ChatGPT 4.1
The gap in capabilities between those models which they tested, and actual current frontier ones is enormous. I would not trust that any conclusions they made are applicable.
It's also worth noting that the very latest models (GPT-6 and Fable 5.1) actually play worse than their immediate predecessors, so it is likely that the labs are not benchmaxxing for this yet. If they did, I'm sure they could come up with something superior to humans. But there is probably very little demand for this compared to IT stuff.
I know HN readers and posters just read numbers and can't be bothered to read, but please read the methodology before making any claims.
> About their ELO ratings from their own website:
> A field-relative rating calculated within ChessBench. It compares performance among the tested models and is not a direct equivalent of a human chess rating.
I am around 1600 elo in over the board I can mop up Astra Fable etc even if I give them literal infinite time and all the subagents and internet access..
Please folks at least use your AIs to read stuff before making claims.
AI is not GM level, it's not even 1600, I am 1600 by using memorized openings people frequently fall for with very basic intuitions.
A GM is 2600 they can beat me in under 20 moves...
Why do I even scroll through this website. For a moment I truly felt fooled, but then I read like a human should.
Maybe I should stop doing that will be a happier life, don't think just believe in the AGI.
Even if you take that website at face value, the ELO scores shown are relative to the other AI models tested, and not comparable to the ELO scores of humans who play against other humans.
I wonder why they didn’t throw a real chess engine in there for a baseline. There are engines where you can set the elo in the settings, so it should possible to see these LLMs relative to a human 1500 rather than just relative to each other.
These ratings seems very wrong, i have beaten GPT Astra max thinking in chess and my rating is close to 1500. The ratings here seem more accurate: https://chessbenchllm.onrender.com/
GPT-6 almost never suggests an illegal move anymore while even Sol still did so time to time
Probably tells us that without labs explicitly training/tuning the models or designing the harness (with fast oracle) the LLMs aren't going to get good at those areas.
1. It’s hard to trust a 2026 paper that’s showing results for such old models.
2. Chess seems to be a poor benchmark for generalized strategic reasoning. People who are good at it rely more on experience and deep domain expertise than on skills that generalize to make them experts at unrelated tasks.
3. The study sounds like proving humans will never fly because they don’t have wings. In reality, humans do fly, and Claude Fable would destroy any human at chess by coding a strong enough engine on the fly.
> People who are good at it rely more on experience and deep domain expertise
People are good are 1900 or 2100 above and the top ones who spend decades in the field i.e. deep expertise are well in the 2200-2700 range.
A 1100 player is none of these things, they are purely relying on strategic reasoning there is a good chance they cannot name a single opening or articulate clearly why a move was appropriate. 1100 is quite low bar.
This isn't how intelligence works. The LLM may not be able to play chess directly through inference, but it can write a program to do it and execute that program. Same as how human intelligence works. We can't fly, but we can build planes.
I think the line of criticism around LLMs sucking at chess makes more sense when you understand what the AI companies are saying about the future trajectory of these models.
The entire recursive self improvement story falls apart once you point out that there is not much "cross domain transfer learning". Meaning that training an LLM to become good at coding, math, etc, will eventually transfer into them being good at other skills that were not explicitly trained for.
Using games like chess which have little economic value is actually a good test for this. What's even more surprising about chess is how much information about chess strategy exists in the training data.
There’s multiple databases of games in algebraic notation. You can also, very easily rl train on pitting models against one another, even without mcts.
It is even worse.. This is a classical reinforcement problem where data generation is easy because the rule set is pre-defined. So you really don't even need any data to start with (but would help).
There are more possible game combinations than atoms in the universe, even those generation of valid game states are as you say pre-defined. that is why models cannot go this route and therefore are poor at chess
Frontier labs don't care about chess. If OpenAI cared, GPT-7 could be a grandmaster+ level chess player. In fact there's a google paper on grandmaster level chess without search with a 270M transformer. Outside that, there was gpt-3.5-turbo instruct which was incidentally a 1800 lichess elo player that didn't make any illegal moves even after a few thousand moves.
Frontier labs care deeply about automating knowledge work and computer use. They are working hard on getting models better and better, and they are succeeding. Astra is a step change on that front. So good luck i guess, if chess performance is your barometer.
> Frontier labs don't care about chess. If OpenAI cared, GPT-7 could be a grandmaster+ level chess player.
If the models were actually intelligent, the way that the boosters claim, they wouldn't need to be tuned to play chess in order to be good at it. That's kind of the point of intelligence, that it is generically applicable to whichever task one wishes.
Pretty much this. Feed it a book or two on chess, and you should have a decent (or good) player. That's the generic intelligence people have. The aims is not to be supremely talented at something, but being able to read a manual and figure how to use/play something. Mastery can be gained overtime.
> the frontier labs are priced according to the narrative that they have produced or will in the very near future produce a fully automated drop-in replacement for most knowledge workers...
Even assuming this is how the AI companies are being valued (they're not), the numbers are off.
The "value" of most knowledge workers -- based on what enterprises currently pay for them -- is $50 - 70 trillion annually. It's reasonable to assume that if AI drop-in-replaced all those knowledge workers, AI companies could credibly charge somewhere in that order of magnitude, because that's what the market is already bearing.
So if their hypothetical revenues are double-digit trillions and valuations are some multiple of that, the entire AI industry would be valued at double-digit trillions at the least.
Yet cumulatively the industry (the frontier labs + the SWAG estimate of the AI parts of all the other players) are valued at, say, ~6 - 7 trillion? Which seems like a fair approximation of how much knowledge work they can currently automate.
Short and to the point! Open and cheap models will undercut the big labs continuously. The blast radius won't be pretty once spending commitments knock the door.
> the best alternative to rigorous specification is human review. human review doesn't scale well to the volumes of output produced by language models. to make matters worse
when the business model is selling more tokens you get such per serve ice times that lead to “more” thinking, engagement baiting, fluffy narratives, and straight up dark patterns
I don’t know who you’re talking about, even the most bearish people like Gary Marcus and Ed Zitron acknowledge that LLMs are useful in these same cases the OP admits. Gary Marcus is even still a long term AI advocate, he just doesn’t think LLMs are enough and we need more foundational breakthroughs. Zitron says it’s valuable technology but not worth the trillion dollar valuations the frontier labs are claiming.
The lack of temperament is very skewed towards the bulls who have been saying AGI is here, software engineering is solved, mathematics is solved, it’s going to destroy the white collar job market, and it’s going to kill us all for like 5 years now.
Even a lot of the people who think that LLMs are a dead end think that we will soon find something signficantly more powerful, which I find deeply alarming. I don't want to know what my white-collar knowledge work will look like in a decade or 2.
Gary Marcus is an especially puzzling addition. If I recall correctly, he has made statements along the lines that superintelligence this century is more likely than not. If you’re AGI-pilled that might read as bearish, but that is still extremely rapid progress in the grand scheme of things.
I can see current limitations, but how do you expect capabilities to change in the next few years? A repeat of the gain that happened in the last two years feels like it would be significant, even if it took a little more than two years this time around.
They do…? Well, “auto” mode has been default in Claude Code for a couple months now. It’s effectively “safer yolo:” tool calls are inspected by a separate classification system (another smaller LLM, I believe) to approve or deny. And you can always layer on additional sandboxing mechanisms to limit the blast radius deterministically.
> They do…? Well, “auto” mode has been default in Claude Code for a couple months now
They have never shipped "yolo" mode by default. Auto mode is not yolo mode. They trained a task specific model just for ensuring the llm didn't accidentally delete every file from your computer.
That is the point I was making, that auto mode is itself a guardrail on top of the model (and not a perfect one.) auto mode seems to cover merely actions the model could take that are clearly bad, like wiping your disk, using an overly privileged context to complete the task, etc.
I recently tasked a GPT model in Codex with implementing part of a new architecture I'm working on. I gave it a very detailed spec and the code it produced looked pretty reasonable and passed my tests. It even did exceptionally well in my evals, so I excitedly declared victory to a few friends. The next day after more careful review I found that the architecture implementation was totally correct, but the model had slipped a one line change to the observation encoding of the RL environment I was prototyping against. The encoding change made the learning problem essentially trivial; the architecture itself, I later realized, had a major flaw that was revealed by returning to the natural encoding.
This is the type of reward hack that is hard to paper over with easy guardrails like auto mode and even harder to specify out. It's also the type of thing a reasonable human wouldn't do unless they were intentionally trying to deceive you.
I agree with the caveat that it's more like a cracked junior engineer who can manage swarms of interns.
Frontier Labs will probably survive off hype valuations but will serve the important purpose of discovering architectures/techniques that will probably spread through rumors/transfers to the rest of the world.
I think bearish on LLMs for automation, and bullish for LLM+human experts in specific fields, is about the right expectation for current architectures.
Apart from issues with task generalization, or perhaps related to it, is the fact that LLMs have real trouble with timekeeping, and cannot estimate the real world time it will take them to do things very well. This plus the memory issues make dreams of long horizon agents, that could plausibly handle changing specifications, quite implausible with current architectures.
In narrow domains with more deterministic outputs though, this is less of an issue, and we see multiple agents succeed much better.
The fusion of that capacity, with humans in the loop able to better direct such agents and act as their temporal tethers, is where I think the real action will be for a while at least.
> This plus the memory issues make dreams of long horizon agents, that could plausibly handle changing specifications, quite implausible with current architectures.
Any reason why that can't be solved through context management and keep-forward scaffolding?
> the models generalize well only on tasks within a small neighborhood of the specific tasks they've been trained on, and even then with severe caveats. the frontier labs have developed a general recipe to teach models almost any specific task enjoying clearly defined levels of task performance; many tasks are covered in the training data
Is this really any different to how humans learn, it takes a lot of training on one specific task to make a human expert as well?
I was a young child when I learned chess by reading a short book, then practicing with a friend. That is not how LLMs learn. I'm no expert on LLMs, but if you showed a human all chess games and books in all history and then said 'play chess' and they still kept making illegal moves, they would have to have a brain injury.
being a bit more specific: the sample efficiency of humans is orders of magnitude larger for more abstract concepts. the same doesn't hold for memory-intensive tasks though (like any kind of trivia), but that only takes you so far.
Humans wear a lot of hats when the do work. They don't even realize how many. My experience with building real systems using LLMs is that you have to be very explicit about such hats and you don't realize how many are worn until you see edge case after edge case after edge case. Check this. Check that. Check this. Check that. Check check check.
Doesn't really even need to look like it. If you can verify rewards, RLVR will optimize really really well. If you can't... it's a struggle. There are probably fewer fields where you can verify rewards than one might hope.
> There are probably fewer fields where you can verify rewards than one might hope.
2 tasks I've done today that I believe robots are nowhere near being able to do: Cleaning my wardrobe and draining bad fuel out of my generator. As in generic use cases.
great, autonomous LLMs will fail. that's actually perfect. they work amazingly well when we're telling them what to do. no autonomy needed, no destruction of humanity. that's all win
refreshing to see amongst the endless tide of "i haven't written a single piece of code since 2025, llms are so good that they have already replaced everyone" gaslighting
DO you know those ice cream shops that sell 30 different flavours. Everybody likes a different flavour, some people dont even like ice cream and buy nothing.
Some people will complain about the wrong flavours, or missing flavours, or the price, the long lines or maybe it closes early on fridays.
Summarise means different things to different people.
https://arxiv.org/html/2509.24239v4
Researchers asked frontier models to play chess. Have a look at the MAR rates in Table 3. When not explicitly told which moves were legal, no model identified legal moves at a rate better than 80%. Many asked for more illegal moves than legal moves. And even when explicitly told which moves were legal, the models continued to ask for illegal moves. With illegal asks discarded, none of the bots could beat a chess model calibrated to 1100 ELO.
The author of the originating post says that "current frontier models need laborious oversight and guardrails on even the simplest tasks", and he's absolutely correct.
> Gemini 2.5 Pro, O3, Claude Sonnet 3.7 and ChatGPT 4.1
The gap in capabilities between those models which they tested, and actual current frontier ones is enormous. I would not trust that any conclusions they made are applicable.
https://chessbench-ai.github.io/#leaderboard
It's also worth noting that the very latest models (GPT-6 and Fable 5.1) actually play worse than their immediate predecessors, so it is likely that the labs are not benchmaxxing for this yet. If they did, I'm sure they could come up with something superior to humans. But there is probably very little demand for this compared to IT stuff.
> About their ELO ratings from their own website:
> A field-relative rating calculated within ChessBench. It compares performance among the tested models and is not a direct equivalent of a human chess rating.
I am around 1600 elo in over the board I can mop up Astra Fable etc even if I give them literal infinite time and all the subagents and internet access..
Please folks at least use your AIs to read stuff before making claims.
AI is not GM level, it's not even 1600, I am 1600 by using memorized openings people frequently fall for with very basic intuitions.
A GM is 2600 they can beat me in under 20 moves...
Why do I even scroll through this website. For a moment I truly felt fooled, but then I read like a human should.
Maybe I should stop doing that will be a happier life, don't think just believe in the AGI.
As a 1500 elo human I can tell you that a 1500 elo chess engine doesn't play like anything like a 1500 elo human.
GPT-6 almost never suggests an illegal move anymore while even Sol still did so time to time
2. Chess seems to be a poor benchmark for generalized strategic reasoning. People who are good at it rely more on experience and deep domain expertise than on skills that generalize to make them experts at unrelated tasks.
3. The study sounds like proving humans will never fly because they don’t have wings. In reality, humans do fly, and Claude Fable would destroy any human at chess by coding a strong enough engine on the fly.
People are good are 1900 or 2100 above and the top ones who spend decades in the field i.e. deep expertise are well in the 2200-2700 range.
A 1100 player is none of these things, they are purely relying on strategic reasoning there is a good chance they cannot name a single opening or articulate clearly why a move was appropriate. 1100 is quite low bar.
The caveat is: It depends on the task.
Are there reams of chess moves that the model can train off of? No.
Are there reams of math papers the model can train off of? Yes.
I think the line of criticism around LLMs sucking at chess makes more sense when you understand what the AI companies are saying about the future trajectory of these models.
The entire recursive self improvement story falls apart once you point out that there is not much "cross domain transfer learning". Meaning that training an LLM to become good at coding, math, etc, will eventually transfer into them being good at other skills that were not explicitly trained for.
Using games like chess which have little economic value is actually a good test for this. What's even more surprising about chess is how much information about chess strategy exists in the training data.
For real??
This is as false as something can possibly be. There are open databases of millions of chess games spanning hundreds of years.
If the models were actually intelligent, the way that the boosters claim, they wouldn't need to be tuned to play chess in order to be good at it. That's kind of the point of intelligence, that it is generically applicable to whichever task one wishes.
> the frontier labs are priced according to the narrative that they have produced or will in the very near future produce a fully automated drop-in replacement for most knowledge workers...
Even assuming this is how the AI companies are being valued (they're not), the numbers are off.
The "value" of most knowledge workers -- based on what enterprises currently pay for them -- is $50 - 70 trillion annually. It's reasonable to assume that if AI drop-in-replaced all those knowledge workers, AI companies could credibly charge somewhere in that order of magnitude, because that's what the market is already bearing.
So if their hypothetical revenues are double-digit trillions and valuations are some multiple of that, the entire AI industry would be valued at double-digit trillions at the least.
Yet cumulatively the industry (the frontier labs + the SWAG estimate of the AI parts of all the other players) are valued at, say, ~6 - 7 trillion? Which seems like a fair approximation of how much knowledge work they can currently automate.
when the business model is selling more tokens you get such per serve ice times that lead to “more” thinking, engagement baiting, fluffy narratives, and straight up dark patterns
The lack of temperament is very skewed towards the bulls who have been saying AGI is here, software engineering is solved, mathematics is solved, it’s going to destroy the white collar job market, and it’s going to kill us all for like 5 years now.
It is literally denialist about current capabilities
They have never shipped "yolo" mode by default. Auto mode is not yolo mode. They trained a task specific model just for ensuring the llm didn't accidentally delete every file from your computer.
I recently tasked a GPT model in Codex with implementing part of a new architecture I'm working on. I gave it a very detailed spec and the code it produced looked pretty reasonable and passed my tests. It even did exceptionally well in my evals, so I excitedly declared victory to a few friends. The next day after more careful review I found that the architecture implementation was totally correct, but the model had slipped a one line change to the observation encoding of the RL environment I was prototyping against. The encoding change made the learning problem essentially trivial; the architecture itself, I later realized, had a major flaw that was revealed by returning to the natural encoding.
This is the type of reward hack that is hard to paper over with easy guardrails like auto mode and even harder to specify out. It's also the type of thing a reasonable human wouldn't do unless they were intentionally trying to deceive you.
Frontier Labs will probably survive off hype valuations but will serve the important purpose of discovering architectures/techniques that will probably spread through rumors/transfers to the rest of the world.
Apart from issues with task generalization, or perhaps related to it, is the fact that LLMs have real trouble with timekeeping, and cannot estimate the real world time it will take them to do things very well. This plus the memory issues make dreams of long horizon agents, that could plausibly handle changing specifications, quite implausible with current architectures.
In narrow domains with more deterministic outputs though, this is less of an issue, and we see multiple agents succeed much better.
The fusion of that capacity, with humans in the loop able to better direct such agents and act as their temporal tethers, is where I think the real action will be for a while at least.
Any reason why that can't be solved through context management and keep-forward scaffolding?
Is this really any different to how humans learn, it takes a lot of training on one specific task to make a human expert as well?
yes.
2 tasks I've done today that I believe robots are nowhere near being able to do: Cleaning my wardrobe and draining bad fuel out of my generator. As in generic use cases.
Some people will complain about the wrong flavours, or missing flavours, or the price, the long lines or maybe it closes early on fridays.
Summarise means different things to different people.