Codex users report higher credit burn as GPT-5.6 makes sequential tool calls
kevinkern · x · 2026-07-27
A discussion is underway about why Codex may be burning through more credits than expected.
One explanation offered: GPT-5.6 often executes tool calls sequentially instead of batching them, because the instructions do not clearly tell it to group independent actions. The result is unnecessary extra usage, and the suggestion is to make the batching behavior explicit so the OpenAI team can see the issue.
More from coding & agent
- Opus 5 code demo improves and the source is now public — bdsqlsz · 2026-07-27
- OpenAI’s Jason uses Codex to triage email, Slack, Linear, and long-running work — petergyang · 2026-07-27
- AI-assisted coding is starting to feel like a multiple-choice exam — dfinke · 2026-07-27
- People doubt Matt Shumer’s game prompt, then run it and build custom games — mattshumer_ · 2026-07-27
- Claude Code Max user spends about $600 in tokens building a game with Matt Shumer’s prompt — mattshumer_ · 2026-07-27
- Grok Build adds a `/deep-research` mode that cross-checks claims before writing the report — XFreeze · 2026-07-27