Which models for which task? 20x Max plan
retained text
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 * Opus 5, Opus 4.8, Opus 4.7, Opus 4.6 * Sonnet 5, Sonnet 4.6 Which ones are folks turning to for feature planning, subagent commanding, code-writing, etc? I'm having mixed luck using Sonnet 5 for all code-writing, because Sonnet 5 is so slow compared to Opus or Fable, but really like how little weekly usage quota Sonnet 5 consumes. I can spawn lots of sonnet instances to build little bits, which is good. But each instance seems to take 10-20 minutes to complete what Opus 5 takes 2-3 minutes or Fable takes 1-2 minutes to complete. Opus 4.8 seems too overly verbose for dialogs about planning, but does a good job at finding bugs or flagging things I didn't consider when describing a feature. I haven't used Opus 4.7 or 4.6 for months to be able to say if they're useful or not. Opus 5 responds quickly and mostly correct, but eats weekly usage quota faster than I'd like Fable is essentially unusable because of how much it consumes weekly usage quota. I've found that for my project, Codex 5.6-Sol pretty much does a better job all-around than all of the claude models, but I'm currently paying for the Claude 20x Max plan, so I don't want to switch providers yet. But when I do end up saying "alright, lemme throw codex at this", codex gets the work done incredibly fast, and accurately, compared to Fable, Opus or sonnet.
document fetched 2026-09-18 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1wizu8v/which_models_for_which_task_20x_max_plan