model roundup

Opus 4.6

13 items · started 2026-09-07 · ongoing (last activity 2026-09-18)

  1. Bring back old Opus in Claude Code Sep 19, 2026 · 1 minutes read Bring Opus 4.6 and 4.8 back to the Claude Code /model picker with the modelPicker setting. I still use Opus 4.6 and 4.8 regularly, as I often find Opus 5 too verbose (and Fab…

  2. Hey Everyone, I'm starting a new role as a remote software developer soon in a great company. I'm a software engineer with several years of experience under my belt and coded several products end to end in different companies in the pre-AI…

  3. I'm looking for some direction on which models everyone is finding are working best for their code-planning/code-implementing/subagent commanding tasks. Of the available models these days that most everyone is using: Fable 5.1, Fable 5 Opu…

  4. In my global instructions, GPT 5.6 Sol and GPT 6 Astra have both spent multiple hours overengineering their own sub-projects that did not support my prompt without narrating a word of what they were working on. Or I will ask a side questio…

  5. Hi there, I'm on the Premium Team plan, but I have had weeks where I chugged my weekly quota in the first half of the week. So I set out to try to optimize my model/effort combo for best output for the least tokens.

  6. It seems that the peak of claude was with Opus 4.6, we used to actually understand what it says and does what was told , sure wasn't the crispiest chip in the bag but got actual job done at an affordable cost and was able to build too many…

  7. As per the title, my girlfriend nor I can subscribe to Max without verifying our IDs. We are in Europe for reference, so GDPR doesn't seem to affect this intrusive KYC.

  8. quick background. i started working with AI (Claude) in February this year with zero coding background before that.

  9. Hello guys, I am Ant, a person can't endure 1 hour long Lidang(Chinese Youtuber) long streaming video, so I made Verbatim. It allows you to only drop 1 YouTube channel link and then it will automatically analyze the whole channel.

  10. Look, everybody has their issues with these models. But I think deep down we all know - Opus 4.6 was the wet dream of AI.

  11. Hello everyone first time posting here. I am a doctor and using claude code for mainly research purposes (writing,statistics,generating and testing hypothesis etc.).I switched to claude from chat when opus 4.6 first released and fell in lo…

  12. Three times in the last twelve months a model made me stop and say whoa. First was Claude Code with Opus 4.6.

  13. Hi all. I have been reading that 4.6 is king, but I always thought that it was related to coding.

← all threads