Just tried it side by side Astra to Opus, Astra was not very eager and Opus crushed it on my one test. I asked it to look at the other hacker news post about speeding up CI, and speed up my CI. Opus made a bunch of good solid changes. Astra just kind of told me this a suggested problem area. I had this issue with 6.1 (sol and astra), its hard to get them to do work. They do well when supervised by Opus though.
The best feature of 6.1 sol was it would actually do stuff for you that Opus refused though. I did use that often.
curious, i haven't experienced what you describe so far. Astra and Sol 6.2 have both gone off and done jsut what I wanted, and as the set of prompts I test against become more vague and handwavy, they both went on to fill in the gaps and produce somehting I was satisfied with.
You and I have the exact opposite requirements. My workflow is that I always like to talk through the problem and the solution first, before codex just starts ripping through the coding portion. I used to have consistent issues getting that kind of workflow. The more recent models from OpenAI are so much better at building a plan collaboratively.