K3 for reference had 0.43 similarity to OpenAI's models, now it's as similar as how OpenAI's models are to each other at 0.58, much higher than other cross-lab similarities.

Comments URL: https://news.ycombinator.com/item?id=50090443

Points: 165

# Comments: 143

- Metaphor overload. We get it, it's just like car racing. Show some mercy on human readers.

- X, not Y

- A, never B

- Tasteless em dashes

- Hallucinated data, like model add date. The standings are so unbelievable that they border on laughable.

I ask it to plan the feature in a separate worktree.

In parallel I start coding without being biased by AI and vice versa.

At some point I read its plan and iterate on it all the while I am in implementation mode. This helps me improve my own vision.

Finally I ask the AI to review my implementation. It flags off bugs and gaps which are usually straightforward for it to fix.

For example model shares

London

Found the time traveller.

Comments URL: https://news.ycombinator.com/item?id=48504672

Points: 10

# Comments: 0

UIs it's generating is pretty good, not without problems, but certainly better than other models at this price point.

For example it's just so natural to share screenshots in a chat.

On my tasks it has not been as good as even Sonnet 4.6 so far.

Instruction following over long context feels worse.

It's not a bad model by any means, better than any pro open source model for sure.

Also as a person developing agentic code tools since before Claude Code, I'm skeptical if str_replace provides accuracy improvement over just full rewrite.

Back in the day when SOTA models would do lazy coding like `// ... rest of the code ...`, full rewrite wasn't easy. Search/replace was fast, efficient and without the lazy coding. However, it came with slight accuracy drop.

Today that accuracy drop might be minimal/absent, but I'm not sure if it could lead to improvements like preventing doc corruption.

My hypothesis is that some models err towards assuming human queries are real and consistent and not out there to break them.

This comes in real handy in coding agents because queries are sometimes gibberish till the models actually fetch the code files, then they make sense. Asking clarification immediately breaks agentic flows.

With search-replace you could work on separate part of a file independently with the LLM. Not to mention with each edit all lines below are shifted so you now need to provide LLM with the whole content.

Have you tested followup edits on the same files?

I couldn't decide on folding and reducing noise so I'm stuck on that front. I believe there is some elegant solution that I'm missing, hope to see your take.

GPT models can follow tool format correctly but don't keep on going.

Grok-4+ are decent but with issues in longer chats.

Kimi 2.5 has issues with it reverting to its RL tool format.

In my personal benchmark it's bad. So far the benchmark has been a really good indicator of instruction following and agentic behaviour in general.

To those who are curious, the benchmark is just the ability of model to follow a custom tool calling format. I ask it to using coding tasks using chat.md [1] + mcps. And so far it's just not able to follow it at all.

[1] https://github.com/rusiaaman/chat.md