I open-sourced a macro execution layer to reduce coding-agent turns (60-task benchmark)

A developer open-sourced Tura, a macro execution layer for coding agents that reduces the number of model turns by executing routine workflow steps together. On a 60-task DeepSWE benchmark, the macro approach with backward reasoning achieved an 80% pass rate with 2,017 rounds, compared to Codex CLI Medium's 63.3% pass rate with 3,140 rounds. The project aims to lower orchestration turns without hiding failures, though the developer notes that fewer turns do not guarantee lower end-to-end cost.

Disclosure: I maintain Tura. A coding agent often spends a separate model turn on each part of a routine workflow: inspect the environment, edit package files, patch the implementation, update tests, build, run lint/tests, then inspect Playwright media. Tura experiments with a simpler execution model: describe that sequence as a macro and let the runtime execute the steps together, while still returning structured results to the agent. The goal is to reduce orchestration turns without hiding failures. On our current 60-task DeepSWE set: | Configuration | Passes | Pass rate | Observed tokens | Rounds | |---|---|---|---|---| | Macro + backward reasoning | 48/60 | 80.0% | 229,695,477 | 2,017 | | Macro Direct | 39/60 | 65.0% | 75,108,167 | 969 | | Codex CLI Medium | 38/60 | 63.3% | 333,538,349 | 3,140 | | Codex CLI High | 36/60 | 60.0% | 455,742,296 | 6,074 | The implementation, task data, and benchmark methodology are public: The important caveat is that fewer turns do not automatically mean lower end-to-end cost. Retries, failures, cache behavior, and completed-task rate all belong in the denominator. I would especially appreciate feedback on the benchmark design and cases where macro execution makes debugging worse rather than better.