haowang02/codex-candy-eval
A Python script that batch-tests OpenAI Codex CLI on a candy math problem and reports reasoning tokens and accuracy.
Collecting fresh signals — velocity needs a few days of history.
collecting data…
star history
The script invokes a locally installed Codex CLI multiple times with configurable model and reasoning effort, parses the responses, and checks whether the answer 21 appears. It aggregates reasoning token usage and correctness across runs. It has no third-party dependencies and is intended as a quick evaluation harness for Codex CLI behavior.
Frequently asked
- What is haowang02/codex-candy-eval?
- A Python script that batch-tests OpenAI Codex CLI on a candy math problem and reports reasoning tokens and accuracy.
- Is codex-candy-eval open source?
- Yes — haowang02/codex-candy-eval is an open-source project tracked on heatdrop.
- What language is codex-candy-eval written in?
- haowang02/codex-candy-eval is primarily written in Python.
- How popular is codex-candy-eval?
- haowang02/codex-candy-eval has 1.1k stars on GitHub.
- Where can I find codex-candy-eval?
- haowang02/codex-candy-eval is on GitHub at https://github.com/haowang02/codex-candy-eval.