Dual-model loop: using GPT Pro and Codex to cross-check math papers
jessi_cata · x · 2026-09-09
A practical dual-model workflow for writing math papers:
- Ask GPT Pro to solve a math problem and flag anything unclear;
- Feed the output to Codex to improve the LaTeX paper draft;
- Send the new draft back to GPT Pro to verify accuracy, then have Codex incorporate the feedback;
The core idea is separating generation from review—letting one model write and another audit—then iterating through multiple rounds to raise the rigor of both derivations and prose.
More from coding & agent
- Tencent paper: continuously harder task environments beat co-evolution, +8.6pp on Terminal-Bench — rohanpaul_ai · 2026-09-09
- Open social layer forces AI agents to disclose identity, bans human impersonation — Capable-Chef-7593 · 2026-09-09
- RoboSim: open-source Rust robot-native simulation engine with physics, sensors and Python bindings — rsasaki0109 · 2026-09-09
- instructor 1.17 ships cache isolation fixes, Gemini retries, local PDF support — jxnlco · 2026-09-09
- Dev: Fable 5.1 Ultracode Workflows Deliver Best Quality, Codex Ultra a Waste of Tokens — jarrodwatts · 2026-09-09
- Put per-turn action schemas in the last user message to preserve prompt caching — Low_Bad_6585 · 2026-09-09