Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Qwen3.8 Max now ranked as the best overall model by agentic index

artificialanalysis.ai

Post reply on HN
Print view View original

271–280 of 364 posts

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#271

Post by ethin »

thejosh wrote: post #123
gnull wrote: post #107

Earlier quoted context omitted.

Claude code is just pool quality. They don't make how this thing will behave clear to the user, or give control. They fail at anything that needs an abstraction or model, not just APIs and shell scripts glued together. And "just ask AI" seems to be the default fix. That vibe coding they brag about as if it was a good thing, it shows. Take their notation for describing permissions. The docs are not comprehensive, and…

so many ridiculous "how the fuck did this get through basic QA?" issues with Claude Code. I can't believe how many critical bugs fall through. My favourite one is the bug where Plan mode can execute destructive commands inadvertently. Then all these get closed with `Closing for now — inactive for too long. Please open a new issue if this is still relevant.`. Awesome.

I mean. This is what happens with vibe-coded projects. When there's no actual software engineering going on, I wouldn't expect anything better than this.

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#272

Post by sieabahlpark »

monster_truck wrote: post #237
jjcm wrote: post #166

China has caught up is the main takeaway here. The SOTA models are so close that it's really hard to compare them intelligence wise - you have to get a feel for them yourself and what works for you. What I'm really excited for is the 27B model. 3.6 is still the king of local, and if 3.8 makes the same improvements it could really legitimately make local viable as a default. I'd love to run a perpetual agent on 3.8 th…

Something I don't think many have internalized is that China has been as good or better for quite a while now (long before anyone was pointing distillation fingers) and enough people have finally tried it for themselves that the understanding has reached critical mass and the careful narrative of american companies is collapsing. When I finally put $15 into Deepseek and it beat the brakes off Codex 5.5 on multiple ra…

[dead]

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#273

Post by ethin »

tempest_ wrote: post #95
aenis wrote: post #84

Earlier quoted context omitted.

I managed to lose around $300 in credits I had saved for some emergency /fast sessions the following way: switch to Fable. Work on the design. Downgrade to Opus for the build. If any of other parallel Opus session has /fast enabled it seems to enable it for the newly spawned session by default. Before I knew it, the $300 was gone. I think the bug is now solved, but it was rather unpleasant. I dont ever remember bugs…

I dont love it. Opus 5 is just a token burner. I use fable plan and spawn opus 4.8 workflows which seems to work alright.

For me it's the opposite: I don't have $200 plus to throw at Anthropic every Month, and when I do get to use Fable it rips through my usage credits like there's absolutely no tomorrow.

Of course, the hilarious thing to me is that Anthropic likes to claim that the usage limits are because of resource allocation problems or something like that. Obviously no such issue exists, otherwise they wouldn't allow you to bypass it by just paying a bit more and it would be a hard limit. So usage credits are entirely their way of just screwing you out of more money.

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#274

Post by vunderba »

cyanydeez wrote: post #270
vunderba wrote: post #265

Earlier quoted context omitted.

Oh I do. Headless 128GB RAM machine serving llama.cpp with a number of local models that I use on a daily basis. • Qwen3-VL picks up new images in a NAS, auto captions and adds the text descriptions as a hidden EXIF layer into the image, which is used for fast search and organization in conjunction with a Qdrant vector database. • Gemma3:27b is used for personal translation work (mostly English and Chinese). • Some s…

recommend opencode w/qwen 35B or 27B with MTP. My secret sauce is to use LLAMAcpp's reasoning-budget and reasoning-message that trigger cut off to overthinking with a message that says to either us subagents or compress the context. opencode's dynamic context pruning plugin can get you pretty far into the stratosphere.

> My secret sauce is to use LLAMAcpp's reasoning-budget and reasoning-message that trigger cut off to overthinking with a message that says to either us subagents or compress the context

Thanks for the tip - I like this a lot. I remember having to do a lot of tweaking to curtail Qwen QwQ-32b when it would go down an endless psychotic recursive reasoning loops as part of its "chain of reasoning."

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#275

Post by swat535 »

pessimizer wrote: post #267
cyanydeez wrote: post #257

Earlier quoted context omitted.

I think americans assume when they see a chinese or asian person working at an american business that they "escaped" china as opposed to just being rich enough to go to school abroad. and has little to no bearing on the amount of intelligent going around.

They've been continuously programmed with insane beliefs about China, which is less shocking when you understand what insane beliefs that they've had programmed into them about their neighbors. The world and your neighborhood are full of evil communists and Nazis who are trying to kill you all the time.

I don't think it's all Americans however there is a portion of them who are not able to grasp the world outside of their borders.

I think it's mainly due to poor education many receive and a very controlled media that suppresses information.

It's shocking considering how much money they spend on education compared to other nations.

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#276

Post by gerdesj »

My vague equivalent of the pelican riding a bicycle test (for a local model without internets) is to ask it: "Where is Yeovil"? I don't expect a totally accurate answer for obvious reasons but I do enjoy watching the accuracy improve.

Qwen3.6-27B-FP8 currently espouses (see below), which is not too bad. The directions are a bit mad but the mileage is about right and there is a helicopter manufacturer here and a RNAS (navy not airforce) museum nearby at Yeovilton. Cosford is in Shropshire which is not a million miles away.

I'm not sure what 盆地的 means but the river Yeo is correct ... OK ... "basin like" - again not bad, even if Chinese is not the first language here. The model understands that Yeovil is named after (or vice versa or at least is associated with) a river

Yeovil is the current form of Gifle (Saxon) which I thought meant "bend in a river" but WP is currently saying "fork in a river". My source is a local museum. There is a fork but was it there 2000 odd years ago? My hydrology skills say ... possibly

---------------------------------------------------- Q: where is yeovil:

Yeovil is a town in Somerset, in the South West of England.

It is located roughly:

    25 miles (40 km) south-west of Exeter
    60 miles (100 km) west of Bristol
    140 miles (225 km) west-south-west of London
Yeovil is known for its historic market town center, RAF Museum Cosford (nearby), and as a significant industrial town, particularly during World War II for aircraft manufacturing (including the Wellington bomber). It sits in the盆地的 valley of the River Yeo.

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#277

Post by TheOtherHobbes »

kennywinker wrote: post #238
BeetleB wrote: post #226

Earlier quoted context omitted.

> If the legal system declares the first thief’s theft not theft But they didn't find it. The Big LLM provider accepted guilt and paid a fine. You can argue whether it was a fair amount they paid, but there is no legal precedent that was set. It's still considered theft.

As i understand it, they accepted guilt for downloading stuff illegally. They didn’t accept guilt for incorporating all of human output into their model without consent.

Copyright law only considers illegal ownership of a work, so the crime - or tort - was making/acquiring copies without permission or payment.

Training from copies has been ruled fair use because it's "transformative" and not simply "derivative."

This is obviously debatable, but that's where the debate is at the moment.

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#278

Post by conception »

AustinDev wrote: post #210
doginasuit wrote: post #198

Earlier quoted context omitted.

Another potential takeaway is that the models all gathering around the same point supports the idea that there is a ceiling to LLM capability.

They always all gather around the same spot then that spot moves every 6-9 months. I think the clustering is more likely evidence of distillation. I don't personally think distillation is a bad thing. If the LLM providers can distill all of human output into their models for 'free'. I don't think distilling a model from the output of those models is morally wrong.

If you talk to the Chinese models, even super smart Qwen 3.8, you can tell they are distilled just from the verbal ticks they have. Gemini, ChatGPT and Claude do not sound alike. The Chinese models 100% sound like one of the 3, usually Claude. American models are load bearing for this LLM generation seam.

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#279

Post by esafak »

Cookingboy wrote: post #266
kadushka wrote: post #258

Earlier quoted context omitted.

Opus 5 and 5.6 Sol are definitely not smart enough to do my job. They require constant supervision. So why would I want to switch to even worse model? Even if it's just slightly worse?

>So why would I want to switch to even worse model? There would be no reason to if you are in the privileged position where cost isn't an issue. For the rest of us something that's 95% as good for 20% the price is a hell of a value proposition.

That's why every benchmark should show the Pareto frontier against cost and latency.

Re: Qwen3.8 Max now ranked as the best overall model by agentic index

#280

Post by michelsedgh »

TheOtherHobbes wrote: post #277
kennywinker wrote: post #238

Earlier quoted context omitted.

As i understand it, they accepted guilt for downloading stuff illegally. They didn’t accept guilt for incorporating all of human output into their model without consent.

Copyright law only considers illegal ownership of a work, so the crime - or tort - was making/acquiring copies without permission or payment. Training from copies has been ruled fair use because it's "transformative" and not simply "derivative." This is obviously debatable, but that's where the debate is at the moment.

So basically because they just browsed and used the information that was mostly public on the internet and they didnt copy it, they just learned from it and thats fine. Which makes sense. None of the llms let u copy someones work exactly anyways... makes total sense honestly. So in this case what happens to distilling? Is that also learning or ur trying to get to their actual weights by kind of reverse engineering it? Where would the argument fall there?
Post reply on HN