Claude Code vs Codex: 6,080 vs 1,644 params for 10-digit addition transformer
DimitrisPapail · x · 2026-11-07
Dimitris Papail's experiment "Addition Under Pressure" asked Claude Code and Codex to each train the smallest possible transformer that can do 10-digit addition.
Results diverged sharply: Claude Code produced a 6,080-parameter model, while Codex delivered just 1,644 parameters — roughly a quarter of the size. A concrete head-to-head comparison of coding agents under an explicit model-minimization constraint.
Related event: AdderBoard challenge: smallest transformer for 10-digit addition(2 posts)→
More from coding & agent
- Open DCR on remote MCP servers invites auth-code hijacking: should you allowlist redirect URIs? — After_Pumpkin3803 · 2026-11-07
- Building an incident-response agent: why memory alone can't be trusted for operational warnings — noufal_rizwan · 2026-11-07
- Letting AI stay on the line after customers hang up turned post-call notes into actions — gsuiteautomations · 2026-11-07
- Tsinghua's Coca-Cola compresses agent tool outputs, cutting tokens up to 89% while boosting accuracy — Tsinghua · 2026-11-07
- ComfyUI-freefield puts Freefield's 117 workflows in a sidebar with one-click model downloads — neuregex · 2026-11-07
- How do you stop micromanaging a custom agent harness? A Reddit dev's research-task dilemma — dangerous_inference · 2026-11-07