Coding and Writing Differences Between Pro and Ultra
Afraid-Reflection-82 · reddit · 2026-07-11
After about 14 hours of use, the author is still figuring out the practical differences between the **Pro web version** and **Ultra Codex**, particularly for coding and writing tasks. Observations include: - **Writing tasks**: Feels overall similar to the GPT-5.5 style, still leaning towards outputting green tables; context handling seems better, especially with the automatic invocation of sub-agents. - **Complex long tasks**: For a 33-page request, the generated structure was less than ideal, possibly approaching the capability limit. - **Coding tasks**: Performed exceptionally well when tested with old PRs, even catching issues that Fable missed. - **UI generation**: The Sites feature is powerful, but without design constraints, it selects suboptimal fonts; the author also noticed others reporting the exact same font choice. - **Workflow ideas**: The author plans to test an "image generation → website generation" pipeline. Finally, a comparison was made: both Pro and Ultra ran the same non-coding writing task. Pro was slightly better, though this might be influenced by existing project context, even with memory disabled.
More from coding & agent
- Why vector databases slow AI agents down after constant writes — PrajwalTomar_ · 2026-07-21
- A 13-minute GitHub Copilot video digs into prompt caching — lee_stott · 2026-07-21
- Codex turns out 123 screensavers in one playful batch — intellectronica · 2026-07-21
- SWE-Pruner Pro trims coding-agent context by up to 39% using the agent’s own states — pmttyji · 2026-07-21
- Autoresearch proposes packaging ML runs as studies with questions, analysis, and code diffs — morgymcg · 2026-07-21
- CHAP defines approvals, handoffs, and audit logs for human-agent workflows — DeliveryTechnical199 · 2026-07-21