News for hackers. Stuff that matters.
Re:GPT‑6.2 Astra and Sol by theanonymousone on Tuesday November 17, 2026 @12:54PM (#50535361)
Happy for new Sol. It's my workhorse on company account.

But more than that waiting for new Luna. To me it was the game changer that ended "Tokenocalypse", and is still almost unbeataben in "infinitely using a decent model"

Re:GPT‑6.2 Astra and Sol by SomeonesAccount on Tuesday November 17, 2026 @12:57PM (#50535419)
I find that Luna tends to misinterpret instructions in READMEs and Agents.mds. Maybe it is something specific about my setup, but I found that Terra is the lowest I can comfortably use on its own (Luna can be driven by larger models in the form of a subagent and it works great for that)
Re:GPT‑6.2 Astra and Sol by MrBuddyCasino on Tuesday November 17, 2026 @01:04PM (#50535517)
I used Luna Max a lot, until Deepseek 1017. Then never again.
Re:GPT‑6.2 Astra and Sol by ariwilson on Tuesday November 17, 2026 @01:38PM (#50536096)
Why not Haiku 5.5? It's substantially smarter than Luna (though more expensive - still cheap though).

(personal benchmark) https://fairbake.dev/signature/agents/haiku-5-5 vs https://fairbake.dev/signature/agents/gpt-6-luna

I've found Luna to be very lazy as well (claiming to have run tests/builds when it hasn't).

Re:GPT‑6.2 Astra and Sol by esafak on Tuesday November 17, 2026 @01:45PM (#50536203)
That's why it's so token efficient. It is so lazy that I feel like it's taunting me to do the grunt work. I wonder if there is a benchmark for that.
Re:GPT‑6.2 Astra and Sol by azuanrb on Tuesday November 17, 2026 @02:48PM (#50537136)
> though more expensive

That’s the main difference. It’s smart enough for most of my tasks, but so cheap that I can use it far more than I could before.