Any improvement to the ChatGPT free plan is really nice. It's easy to forget that most people have never used a SOTA model, and only think of their experience with GPT-4o or Google Search's AI Overviews when asked about AI.
Terra is far from awful. I've not used Luna enough to judge whether it's significantly better than Luna, but at least via GitHub Copilot, Terra is leaps ahead of Sonnet 5.
I've had great success with using Luna and having DeepSeek 4 Flash check Luna's output. Oh My Pi has a mode built-in that does this automatically ("advisor" mode). Deepseek only interrupts when it spots an issue so it doesn't slow Luna down.
Both models are cheap enough that I can run 4 sessions at the same time without running out of the 20 USD codex and 10 USD Opencode plan. I've burned through almost a billion tokens this week and I've done some pretty big refactors as well.
I have a Claude Max subscription but I've barely been using it because of the many issues they've had this week.
It used to be 10 messages every 5 hours using GPT-5, then unlimited 4o-mini. More recently, it went down to ~5 messages per day on 5.5 Instant, then unlimited on 5.5 mini.
luna is very good