cd /news/artificial-intelligence/macro-execution-reduced-codex-rounds… · home topics artificial-intelligence article
[ARTICLE · art-66314] src=discuss.huggingface.co ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

Macro execution reduced Codex rounds by up to 84% in a 60-task benchmark

A macro execution approach reduced Codex rounds by up to 84% in a 60-task benchmark, according to Tura maintainer and open-source implementer. Testing Codex CLI 0.144.1 with GPT-5.6-sol on 60 DeepSWE debugging and rewrite sessions, the Macro Direct configuration used 83.5% fewer observed tokens and 84.0% fewer rounds than Codex CLI High, while achieving a 63.3% pass rate versus 60.0%. The Macro + backward reasoning configuration reached an 80.0% pass rate with 2,017 rounds and $221.14 estimated cost.

read1 min views4 publishedJul 21, 2026

I maintain Tura and open-sourced this implementation. This is a project benchmark, not an independent product review.

A typical repository task can spend separate model turns on environment inspection, several edits, building, tests/lint, and reading Playwright media. The macro approach defines that deterministic workflow once, runs it below the model boundary, and returns the combined evidence when another reasoning decision is actually required.

I tested Codex CLI 0.144.1 with GPT-5.6-sol on 60 DeepSWE debugging and full-repository rewrite sessions:

Configuration Passes Pass rate Observed tokens Rounds Estimated cost
Macro + backward reasoning 48/60 80.0% 229,695,477 2,017 $221.138
Macro Direct 39/60 65.0% 75,108,167 969 $99.620
Codex CLI Medium 38/60 63.3% 333,538,349 3,140 $257.173
Codex CLI High 36/60 60.0% 455,742,296 6,074 $327.483

Against Codex CLI High, Macro Direct used 83.5% fewer observed tokens and 84.0% fewer rounds, with a 5-point higher pass rate. The backward-reasoning configuration spent part of the saved budget on stronger investigation and verification and reached 80.0%.

This is not blind batching: the macro stops whenever new evidence requires a decision. I also would not generalize the aggregate table to every repository; task mix, failure handling, and macro boundaries matter.

Implementation, scripts, and data:

Benchmark record:

Where would you draw the boundary between a safe deterministic batch and a step that should always return control to the model?

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @codex cli 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/macro-execution-redu…] indexed:0 read:1min 2026-07-21 ·