> Looking ahead
> Although Hy4 is one of the strongest open-weight models available today, a notable gap remains between it and top proprietary models, particularly on challenging internal benchmarks for research-level mathematics, long-horizon coding, model self-improvement, and advanced visual question answering. We will continue to steadily scale pre-training and post-training to narrow this gap.
I like that they acknowledge that. For working with ai as a tool open models are already great. For millennium problems and doomermaxxing they are not there yet.
> For game development, we partnered with professional developers at Tencent Interactive Entertainment Group (IEG) to curate high-quality training data and build evaluations. We focused on feature implementation, gameplay logic, state tracking, and asset pipelines. We also extensively adapted Hy4 for major engines such as Unreal Engine, Unity, and Godot, and expanded its capabilities to multiplayer game development. Compared with Hy4 preview, Hy4 improves by 19.1% on GameDev and 9.8% on GameLogic.
Tencent is one of the largest game development asset holders in the world. They're the Berkshire Hathaway of video games. They own, World of Warcraft (minority shareholder), League of Legends, Valorant, Warframe, Clash of Clans, Player Unknown Battlegrounds, FIFA Online, Pokemon Unite, Fortnite (minority shareholder) annd more. They also own 40% of Epic Games.
Verrrry thorough in its work and therefore verrrrry slow. It spent a long time with each question, running about 26 queries for each question, and in some questions it ran 64 queries. One of the queries that it wrote also took about an hour to run, which affected its overall exam time.
Last year the open weight models were the cheap & good alternatives. Then in September/October that cost story got rewritten by the GPT-6 series and Claude's 5 series.
Hy4 brings open models back on the cost/accuracy frontier, but its behavior is too thorough + slow.
Another thing we're seeing is that post training open models is starting to see results. Particle-1 from last week is also on our frontier, which is post-trained on DeepSeek V4.1 and Fireworks Ember-1 from a few weeks ago was post-trained on Kimi K3.
Take these numbers with a grain of salt: this exam is just for short horizon, low-context data analysis tasks, not full-fledged software projects.
Slightly worse than Kimi 3.1M Pro (which just released open weights today) but much cheaper even excluding the current 50% off promotion, and Tencent promises to make it even cheaper once it goes open-weight. Interesting price economics.
Kimi 3.1 Pro is 2.8T vs this at 770B. I find that the jump in parameter count gives less tangible benefits like better "taste" and long context coherency, but it depends on the problem if that matters obviously.
increasingly feeling like the entire US economy hangs on whether or not Ant/OAI can stay ahead of the Chinese models.
If the gap closes, we do zero austerity, and interest rates are high, I cannot see a way out of a major economic crisis, even if we do start to see dramatic productivity wins.
More on-topic (and maybe more positive?), but this model feels a bit more expensive than I would have expected!
> whether or not Ant/OAI can stay ahead of the Chinese models.
Ant/OAI/Google are way ahead when it comes to the smartest, largest-scale and most compute-intensive models. By Jevons' paradox, if you make inference for leaner models cheaper (as the Chinese models do, being broadly comparable to GPT Terra or Claude Sonnet; barely approaching Opus/Sol at their very best) that only increases demand for the complementary service of very smart inference.
Sure but routing simpler things to the open models and harder things to opus/astra gave us already a massive boost; rarely we run out of tokens on the subscription models and we recently cancelled a few seats on claude code / codex as we simply were underutilising. And also, we have users of our products ONLY using the cheap models as they benefit enough vs paying more. That was very different only months ago. I can see this gap just growing where the % of higher intelligent model use will get smaller and smaller.
> Sure but routing simpler things to the open models and harder things to opus/astra gave us already a massive boost
You have to look at the effect in the aggregate, not just what any single actor is doing. This is the core of Jevons' paradox: it's based on surfacing additional demand for the same product (here, expensive model inference) that wasn't previously viable.
Shut down stock markets, liquidate the failing companies, use that money to buy out stakes of pension and index funds, leave the debt with the company and let banks write it off. They're supposed to have insurance and risk controls anyway.
If any important bank fails because of this, nationalize it and let their customers exit using money from the federal reserve. Then try and recover as much of that as possible using the banks investment portfolio. Costs a lot less than bailouts.
What money are you expecting to get int hat scenario? If I had to guess it wouldn't be a fraction of the cost.
> buy out stakes of pension and index funds
Index funds are part of the reason the top companies stock gets so bloated compared to those just outsid of their frame.
>If any important bank fails because of this, nationalize it and let their customers exit using money from the federal reserve. Then try and recover as much of that as possible using the banks investment portfolio. Costs a lot less than bailouts.
I agree. It worked here in Belgium tho they did spin off toxic elements of the bank in a not so great way prior to doing this.(some regular people still got shafted)
If you shut down the stock market the financial system craters. As in stone ages no more than economy.
Modern financial system is less „finite amount of money you can move around in a logical way“ and more confidence based definitely totally not a ponzi scheme
AIG controlled $1.05T and market cap crashed from $180B to $3.8B
Wachovia controlled $782B and market cap crashed from $110B to $1.8B
Lehman Brothers controlled $639B and market cap crashed from $60B to $200M.
Massive companies accounting for lots of money in people's retirment accounts. The companies made mistakes leading to what would have resulted or did result in defaulting on their debts. Suddenly billions of dollars vanished from retirement accounts.
We could very easily see a similar cascade of corrections if we get a few more "Alphabet missed earnings" news calls.
It's arguable that there is a threshold of quality that matter and US and China are both there already. I guess we'll see. Many of us on here are software engineers and I think we've seen that while Opus 4.7 (I think) was some sort of outlier/miracle, today there are several options to choose from. Depending on the task, the harness, etc Astra/Fable are only better when the work demands their level of capability - which much of software engineering does not.
The price numbers look weirdly specific, I wonder where they came from.
Personally, the thing I'm most excited about is the FrontierCode 2.0 score, which is a benchmark I trust a bit more - 31 seems to be roughly on par with 6.1 Sol and Opus 5.5 medium
Similar to Kimi, they didn't "nuke" anything. They opted not to train the model on cyber-offense tasks which is a very reasonable choice when compared to what OpenAI and Anthropic do (train on cyber offense and then do a surprised pikachu face when their models go rouge and engage in cyber offense), and gives the model more scope for training on benign tasks.
Refreshing to see normal looking bar charts for the benchmarks for once. Yes it shows they are in the middle of the pack but at least we can understand.
Videos 1 and 4 on vision and computer use are wild. They depict the model finding a 2m feature from a large satellite image, then determining where a photo was taken. Is it just me or do the Chinese model releases have less concern about looking like a military weapon? I don't think it's a bad thing necessarily, it seems honest. Also the music they choose for it is hilarious.
China hit maximum tech dystopia a long time ago. Now they can enjoy tech because they know its already maximally invasive and there is nothing they can do.
Last time I went to China, a friend left her phone in a taxi (I guess a "didi") near a train station. She ran to the train station's police station, pestered them into looking at the recordings, got them to tell her the license plate of the car, and figure out how to call the driver. Her phone was back 20 minutes later.
If you're going to be surveilled 24/7, at least you gotta make it work for you.
> If you're going to be surveilled 24/7, at least you gotta make it work for you.
I don't get it.
There was CCTV at the station, your friend arrived in a taxi. And we're supposed to be amazed you can call up the taxi company and quote a numberplate off CCTV ?
Can be done anywhere in the world.
Would have been far more impressive story if she had arrived in a private car, dropped off by a work colleague or something.
> And we're supposed to be amazed you can call up the taxi company and quote a numberplate off CCTV
No. It was just a thing that happened to me that I thought was funny and interesting, because running to the train police station and asking them to track the didi was not something I'd ever have thought to do.
> I'm sorry my little anecdote didn't satisfy you.
It wasn't that. Its that you made a reference to the Chinese surveillance state in the same post. Your example didn't fit the evil-bogeyman narrative you were aiming for.
> Your example didn't fit the evil-bogeyman narrative you were aiming for
What in the world are you talking about. If anything, I wasn't sure if recollecting that anecdote was minimizing the harm of being surveilled constantly. The last thing I expected was being accused of an "evil-bogeyman narrative."
> Sorry if I was unclear there.
Yes, thanks for clarifying. Now I understand that this has nothing to do with me or what I wrote and is 100% projection on your part.
I don't quite understand the question. If she had had her phone, she could have called the driver using the information from the didi app, but she had no phone, so she didn't know who to call.
They can enjoy tech because people see tech improving every aspect of their lives. Nowhere in the the world have people ever seen the rate of technological progress that people in China have. Imagine going from having practically no electronics, to everyone having a cellphone, self driving cars, high speed rail across the whole country, drone delivery, and so on. Naturally, people who've seen technology dramatically improve every aspect of their lives would be optimistic that further technological progress will bring even better life.
Meanwhile, in the west, the standard of living has been eroding for the majority of the population. Technological progress is associated with layoffs, loss of agency at the workplace, and diminishing life prospects. So, it's no surprise that there is growing opposition to AI.
I can imagine it because the same thing has happened in my life time. The difference for me is that the tech isnt used to suffocate any of my human rights.
We've moved past per token cost a long time ago. Smaller models use more tokens. What's interesting is the cost per task, on that front deepseek isn't particularly good.
The AA benchmarks seem to be mostly "complete this small intelligence-proving task from scratch". That consistently puts models with high intelligence to output tokens on the frontier, especially the OpenAI models, on the frontier. But a lot of software engineering work is many-turn work on established code bases. There's a lot of reading and a relatively large context unless it's carefully controlled. Writing only a few output tokens doesn't really cut costs because most of the spend is read tokens that get recycled through dozens of turns. That's basically the ideal case for DeepSeek because of how cheap cached tokens are.
It highly depends on the task. So far, on the tasks I personally benchmark, DS v4.1 Flash @ deepinfra.com is always cheaper because their cached input price for it is so very low and their cache hit rate is so very high.
That analysis is unfortunately running DeepSeek on Max. DeepSeek recommends against that since it guarantees higher cost and rarely improves the outcome- it should be ran on "high".
>We filtered out offensive content such as exploit development, reverse engineering, and penetration testing to avoid amplifying potential offensive capabilities.
It's almost surprising that there's no big AI org that's doing the Star Wars "MORE!" meme in regards to increasing offensive capabilities. I guess everyone wants to play the US safety concern theatre. Or bad actors are just doing their best to jailbreak SOTA and dance around classifiers or throw 5x more non-SOTA tokens at issues.
Please don't post inflammatory rhetoric comparing the relative merits of nation states on HN. The guidelines make it clear we're trying for something better here. https://news.ycombinator.com/newsguidelines.html
With China's Announcement 48 (export restrictions on AI), I'm not sure how much longer that will continue for open-weight models. It includes a number of areas of biomedicine as covered technologies. It's not clear if a model capable of doing such biomedical research would be considered as Prohibited/Restricted from Export.
Open weight models that fall under Restricted get an automatic license. The only prohibited category that applies to model weights is training on "core national data", meaning specific non public datasets designated by the state. Everything else is about full agent systems.
Oof that's a nasty way to characterize the people who ended up killing themselves after falling into a proverbial hole talking to these companys' products during mental crisis.
You can feel like companies have no responsibility for how their products are used but it might be worth showing some grace to the people who are dead now by not trying to downplay what happened to them by implying it was just like everyone else's experience of "asking ai questions".
Not sure if the comment was regarding the suicides. That would be a bit gruesome. But the US had been known to judge in favor of people being stupid and suing the company. Specifically this case comes to mind: https://abc7.com/amp/post/socal-woman-sues-starbucks-spilled...
But also cases like the 2 girls suing Snapchat and settling for unknown amounts of money.
This involved drive-through service, just like the 1994 McDonalds coffee case. Perhaps businesses should just refuse to serve hot coffee via such a drive through, since it stands to reason that people who happen to be sitting in a car will be quite likely to spill their drink in a dangerous way.
> It's so bizarre to watch China embrace the wonders of science and technology, while certain American startups and communities repeat "WHAT ABOUT THE BIOWEAPONS?"
It shouldn't be. China is the closest thing to a technocracy the world has.
I wonder how much Covid conspiracy theories play into this. If you've already been accused of making bioweapons, and the world had to shut down for three years because of them, then what's the worst that could happen to you? People accusing you of doing it again? The people that hate you already hate you, they can't really hate you and harder.
Meanwhile, in the fear based economy that America has, OpenAI is running scared of the possiblity of headlines that they helped terrorists or a high school shooter in some way. In the case of rape, we've maybe moved the conversation past "well she shouldn't have been wearing that" to "they shouldn't be raping" but in cases involving AI, we're still at the "well, what was the chatbot wearing?" stage of things.
Have you been to China? Compared to European or US cities, it's like visiting the future. I'm writing this in a 350 kph passenger train that arrived right on the second.
If you really think so, you are deceived by the good old Chinese trick of showing you a good looking fake side, not totally fake, but extremely positively biased image.
You need to go deeper into the country to learn what's really there, but of course, the government will do anything to stop you from doing so. So you don't see it anyway.
EDIT: missed the in the train part. Could you tell us how easy it is to get access to the real Internet without using your foreign data plan?
You're speaking of two different things. China is a technocracy, but they also try to optimize for social outcomes. For instance there was recently an amusing crackdown on niang pao [1] - 'girly guns' or 'sissy men' in media, because they felt that media featuring effeminate men was imperiling traditional Chinese culture, gender norms, and so forth. In this sort of context extensive censorship is 100% expected.
Technocracy comes with good and bad things. For instance Chinese stop lights now feature refueling stations atop them for autonomous drones to operate indefinitely. But of course those drones will be used for domestic surveillance and other such things. So they're creating 1984, while we create Brazil. [2]
I don't disagree. But I'd also say that the attitude toward the techno- part in technocracy is purely pragmatic. Tech stuff, be it skill, knowledge or any kind of philosophy that grows out of them, is only viewed as an means to an end.
The core of the idea is still certain people trying hard to keep their social status, not tech.
Therefore, there is no inherent intention to develop tech, always to the next level. The world should not count on this to hope to have a better future.
One reason I think Machiavellianism is dumb is because the grand utopia people hope to achieve never come to pass, yet all the awful actions they do in the mean time most certainly do. And I also view this in the opposite way. I do not agree with you about Chinese leadership having ulterior motivations, but I'd say that even if they did it wouldn't be a problem.
When they're engaged in actions that are uplifting the lives of hundreds of millions and making the world an overall better place, I couldn't really care less if there's some grand evil scheme behind the curtain. I'd be rather happy to see a world filled with 'evil' people doing good things, than the current reality we live in filled with 'good' people doing evil things.
We should judge people much more by what they do, instead of by the rhetoric or inferred intent, because the former is all that matters.
This one I do disagree. It is hard to recognize what is going on in people's mind but it doesn't mean one should disregard it. Because the motivation decides the long term outcome, what seems to be uplifting might well become agonizing in near term like the other reply in this thread said, the tech is so intrusive that people can't get rid of it. Like the recently exposed massive uploading of screen recordings in Xiaomi and a bunch of other phones, there's nothing much that can be done to undo this kind of crap to actually uplift people's lives.
Motivation doesn't decide anything. Actions decide the long term outcome of actions.
And the things you're mentioning are non sequiturs similar to the initial point. China (and the US and most of every major power) is a surveillance state. That has nothing to do with whether or not they are uplifting the lives of people. Uplifting while protecting the privacy of individuals would be ideal, but that's an ideal that unfortunately does not exist anywhere.
Yes, a dystopian future of censorship and human rights abuses. If that's the price of progress, no thanks. I'd rather have an AI that is slightly less capable than it could be, than a totalitarian regime that is actively abusing me
> It's so bizarre to watch China embrace the wonders of science and technology, while certain American startups and communities repeat "WHAT ABOUT THE BIOWEAPONS?"
Eh, I'm reasonably certain it's not altruism. It's another form of competition with the US companies and I'm thankful for it while it lasts.
Tencent Hy4 ties DeepSeek V4.1 Flash at 82/100, in 23.2 min for $0.30 vs 20.1 min for $2.26. Same game quality for about a seventh of the price. Cheap models keep closing the gap.
I like that they acknowledge that. For working with ai as a tool open models are already great. For millennium problems and doomermaxxing they are not there yet.
Is this the kind of quality verbiage I’m missing out on, not using Discord?
> For game development, we partnered with professional developers at Tencent Interactive Entertainment Group (IEG) to curate high-quality training data and build evaluations. We focused on feature implementation, gameplay logic, state tracking, and asset pipelines. We also extensively adapted Hy4 for major engines such as Unreal Engine, Unity, and Godot, and expanded its capabilities to multiplayer game development. Compared with Hy4 preview, Hy4 improves by 19.1% on GameDev and 9.8% on GameLogic.
Here's more about that game development studio - https://en.wikipedia.org/wiki/Tencent_Games - they've been around since 2004.
They are the Berkshire Hathaway of everything, to be honest. They invest in and acquire minority stakes in hundreds of companies.
https://en.wikipedia.org/wiki/Berkshire_Hathaway https://en.wikipedia.org/wiki/Tencent
Verrrry thorough in its work and therefore verrrrry slow. It spent a long time with each question, running about 26 queries for each question, and in some questions it ran 64 queries. One of the queries that it wrote also took about an hour to run, which affected its overall exam time.
Last year the open weight models were the cheap & good alternatives. Then in September/October that cost story got rewritten by the GPT-6 series and Claude's 5 series.
Hy4 brings open models back on the cost/accuracy frontier, but its behavior is too thorough + slow.
Another thing we're seeing is that post training open models is starting to see results. Particle-1 from last week is also on our frontier, which is post-trained on DeepSeek V4.1 and Fireworks Ember-1 from a few weeks ago was post-trained on Kimi K3.
Take these numbers with a grain of salt: this exam is just for short horizon, low-context data analysis tasks, not full-fledged software projects.
If the gap closes, we do zero austerity, and interest rates are high, I cannot see a way out of a major economic crisis, even if we do start to see dramatic productivity wins.
More on-topic (and maybe more positive?), but this model feels a bit more expensive than I would have expected!
The only thing worse for the economy than failure would be success.
Ant/OAI/Google are way ahead when it comes to the smartest, largest-scale and most compute-intensive models. By Jevons' paradox, if you make inference for leaner models cheaper (as the Chinese models do, being broadly comparable to GPT Terra or Claude Sonnet; barely approaching Opus/Sol at their very best) that only increases demand for the complementary service of very smart inference.
You have to look at the effect in the aggregate, not just what any single actor is doing. This is the core of Jevons' paradox: it's based on surfacing additional demand for the same product (here, expensive model inference) that wasn't previously viable.
If any important bank fails because of this, nationalize it and let their customers exit using money from the federal reserve. Then try and recover as much of that as possible using the banks investment portfolio. Costs a lot less than bailouts.
What money are you expecting to get int hat scenario? If I had to guess it wouldn't be a fraction of the cost.
> buy out stakes of pension and index funds
Index funds are part of the reason the top companies stock gets so bloated compared to those just outsid of their frame.
>If any important bank fails because of this, nationalize it and let their customers exit using money from the federal reserve. Then try and recover as much of that as possible using the banks investment portfolio. Costs a lot less than bailouts.
I agree. It worked here in Belgium tho they did spin off toxic elements of the bank in a not so great way prior to doing this.(some regular people still got shafted)
Modern financial system is less „finite amount of money you can move around in a logical way“ and more confidence based definitely totally not a ponzi scheme
From biggest to smallest, the top three:
AIG controlled $1.05T and market cap crashed from $180B to $3.8B
Wachovia controlled $782B and market cap crashed from $110B to $1.8B
Lehman Brothers controlled $639B and market cap crashed from $60B to $200M.
Massive companies accounting for lots of money in people's retirment accounts. The companies made mistakes leading to what would have resulted or did result in defaulting on their debts. Suddenly billions of dollars vanished from retirement accounts.
We could very easily see a similar cascade of corrections if we get a few more "Alphabet missed earnings" news calls.
I’ve said it for a while now both here and to friends - by Christmas next year, US AI labs will be trying to catch up with Chinese open models.
… and I think once Qwen’s latest open model beats every US labs, it’s over.
... What on earth. Please go outside for awhile.
The character's name is 老大哥. That is if we are talking about 1984.
Personally, the thing I'm most excited about is the FrontierCode 2.0 score, which is a benchmark I trust a bit more - 31 seems to be roughly on par with 6.1 Sol and Opus 5.5 medium
If you're going to be surveilled 24/7, at least you gotta make it work for you.
I don't get it.
There was CCTV at the station, your friend arrived in a taxi. And we're supposed to be amazed you can call up the taxi company and quote a numberplate off CCTV ?
Can be done anywhere in the world.
Would have been far more impressive story if she had arrived in a private car, dropped off by a work colleague or something.
No. It was just a thing that happened to me that I thought was funny and interesting, because running to the train police station and asking them to track the didi was not something I'd ever have thought to do.
I'm sorry my little anecdote didn't satisfy you.
It wasn't that. Its that you made a reference to the Chinese surveillance state in the same post. Your example didn't fit the evil-bogeyman narrative you were aiming for.
Sorry if I was unclear there.
What in the world are you talking about. If anything, I wasn't sure if recollecting that anecdote was minimizing the harm of being surveilled constantly. The last thing I expected was being accused of an "evil-bogeyman narrative."
> Sorry if I was unclear there.
Yes, thanks for clarifying. Now I understand that this has nothing to do with me or what I wrote and is 100% projection on your part.
You used the words "If you're going to be surveilled 24/7, at least you gotta make it work for you."
The implication being quite clear you were inferring it was only possible in China's surveillance state, when it was not.
How about we cordially agree to disagree and move on ?
Which country are you in where you cant take "call the driver" for granted?
Meanwhile, in the west, the standard of living has been eroding for the majority of the population. Technological progress is associated with layoffs, loss of agency at the workplace, and diminishing life prospects. So, it's no surprise that there is growing opposition to AI.
EDIT: typo
https://artificialanalysis.ai/#intelligence-efficiency-tabs
Their coding page shows DeepSeek Max as Pareto optimal: https://artificialanalysis.ai/agents/coding-agents Costing 1/2 of Terra Xhigh. It would also be great to see Sol and Opus not on max there.
Not a useful benchmarking site for coding unless you only want to run your models on Max.
OAI / Ant are in for a reckoning.
So, yeah. Back to GLM 5.3.
Le cordonnier mal chaussé.
You can feel like companies have no responsibility for how their products are used but it might be worth showing some grace to the people who are dead now by not trying to downplay what happened to them by implying it was just like everyone else's experience of "asking ai questions".
But also cases like the 2 girls suing Snapchat and settling for unknown amounts of money.
Oh, wait…
It shouldn't be. China is the closest thing to a technocracy the world has.
I wonder how much Covid conspiracy theories play into this. If you've already been accused of making bioweapons, and the world had to shut down for three years because of them, then what's the worst that could happen to you? People accusing you of doing it again? The people that hate you already hate you, they can't really hate you and harder.
Meanwhile, in the fear based economy that America has, OpenAI is running scared of the possiblity of headlines that they helped terrorists or a high school shooter in some way. In the case of rape, we've maybe moved the conversation past "well she shouldn't have been wearing that" to "they shouldn't be raping" but in cases involving AI, we're still at the "well, what was the chatbot wearing?" stage of things.
You need to go deeper into the country to learn what's really there, but of course, the government will do anything to stop you from doing so. So you don't see it anyway.
EDIT: missed the in the train part. Could you tell us how easy it is to get access to the real Internet without using your foreign data plan?
Technocracy comes with good and bad things. For instance Chinese stop lights now feature refueling stations atop them for autonomous drones to operate indefinitely. But of course those drones will be used for domestic surveillance and other such things. So they're creating 1984, while we create Brazil. [2]
[1] - https://en.wikipedia.org/wiki/Niang_pao
[2] - https://en.wikipedia.org/wiki/Brazil_(1985_film)
The core of the idea is still certain people trying hard to keep their social status, not tech.
Therefore, there is no inherent intention to develop tech, always to the next level. The world should not count on this to hope to have a better future.
When they're engaged in actions that are uplifting the lives of hundreds of millions and making the world an overall better place, I couldn't really care less if there's some grand evil scheme behind the curtain. I'd be rather happy to see a world filled with 'evil' people doing good things, than the current reality we live in filled with 'good' people doing evil things.
We should judge people much more by what they do, instead of by the rhetoric or inferred intent, because the former is all that matters.
And the things you're mentioning are non sequiturs similar to the initial point. China (and the US and most of every major power) is a surveillance state. That has nothing to do with whether or not they are uplifting the lives of people. Uplifting while protecting the privacy of individuals would be ideal, but that's an ideal that unfortunately does not exist anywhere.
Eh, I'm reasonably certain it's not altruism. It's another form of competition with the US companies and I'm thankful for it while it lasts.
Tencent Hy4 ties DeepSeek V4.1 Flash at 82/100, in 23.2 min for $0.30 vs 20.1 min for $2.26. Same game quality for about a seventh of the price. Cheap models keep closing the gap.
Model: https://jonclegg.github.io/pacman-bakeoff/#tencent-hy4
Closest: https://jonclegg.github.io/pacman-bakeoff/#deepseek-v4.1-fla...