mims

Agentic Software Dev | QuorX Agentic Harness

[Cline @cline]

Kimi K3.1 Pro nearly matches Opus 5.5 performance on Terminal-Bench 4.0 at *half* the cost.

Moonshot also removed cyber data from mid-training and RL to deliberately reduce performance on exploit tasks by ~10%, while everything else, including coding, kept improving.

[Chubby♨️ @kimmonismus]

Holy shit: Kimi k3.1 pro is on par or even outperforms Opus 5.5 and GPT-6 astra on key benchmarks!!

They just announced K3.1 Pro, a 2.8T-parameter multimodal model with a 1M-token context window that it claims rivals leading proprietary models in coding and agent tasks, with open weights promised by November 4.

China is back!!

[mims @mims3590]

@giadotai Lmao true it always will be threeJS xd .. but your right it would be amazing 3d production assets

[mims @mims3590]

@KATEWWANG congratz keep it going

[mims @mims3590]

@sophie_launch https://t.co/uxX0hIwWvS

[mims @mims3590]

@sagarmoy22 @X I use to be into Ts/Js related fields now seeking into go, What about you

[mims @mims3590]

@PickOrange amazing

[mims @mims3590]

@Shri_73_ Dev-to-dev then. I write code, break it, fix it, repeat. What's your DSA grind looking like right now? Mine's mostly https://t.co/I3tur7X21j

[mims @mims3590]

@brian_millot https://t.co/wzNZVzWbVC

[mims @mims3590]

We have credited all sources at the end of the article. Do check them out as well

[mims @mims3590]

How to direct a video using video gen or just code ?
Here is a article deep dive with visuals we found in our research when developing skills for QuorX Ai harness

https://t.co/onNBoP8AMZ

[mims @mims3590]

@lightsilver323 https://t.co/uxX0hIwWvS

[mims @mims3590]

@SamVale29 https://t.co/uxX0hIwWvS

[mims @mims3590]

@SamVale29 https://t.co/uxX0hIwWvS

[mims @mims3590]

Using a brand new chat every single time makes it almost impossible to get close to this. Just my take—but I've hit that wall hard when I stuck to the same chat or ran multiple agents on high Opus.

When I ran multiple agents in separate new chats on medium though? More usage, roughly the same results, way more work done.

[mims @mims3590]

@KATEWWANG Lets add more to that count
Lets connect am into Software Dev
Check me out https://t.co/dIYMwGp9Nz

[mims @mims3590]

@elder_plinius Needed this way too much negativity came out there indirectly

[mims @mims3590]

Hmm seems like @grok 4.6 Extra High now have 500k context length option in grok build

Nothing big deal but this model is really good at overall task due to low hallucination and moderate token use for its time

[mims @mims3590]

@VeraYugen @X Let's connect am Software Developer
https://t.co/dIYMwGp9Nz