# Read a coding-agent result

Start with the task, then each measure.

Code-season results are pending. No leader is established on any measure.

Reading order: Task scope → Measure → Evidence.

[Review the reading method](/read-results#measures)

## Check whether the task fits

Small Python tasks; production relevance is unproven.

### Read the scope and limits

Season one concerns small, original Python repositories. Forty pilot tasks precede a separate official corpus. It does not establish performance on production codebases.

## Read each measure separately

A measure is not an overall ranking.

### Read the scope and limits

Read quality, coverage, cost and speed with their evidence and uncertainty. A lead on one measure does not establish an overall champion. An unmeasured dimension stays unmeasured.

## Look for missing alternatives

Observed alternatives, not admitted participants.

### Read the scope and limits

Two independent organizations mention five options: Claude Code, Codex CLI, GitHub Copilot coding agent, Cursor and Gemini CLI. Claude Code, Codex CLI and Cursor appear in both accounts. Mentions are not performance results.

### No coverage claim follows yet

This observed consideration set was not frozen before the rights review. It is a candidate, not an admitted roster or the denominator of a published coverage claim. The Copilot account concerns its cloud coding agent, not local CLI adoption.

## Follow the evidence

Keep each claim within its evidence.

### Read the scope and limits

Check configurations, dates, corpus, frozen rules and missing evidence. Historical examples belong to their own categories. They do not establish a coding-agent result.

### Reuse a published result

For a published result, keep its supplied citation, license and attribution together. The citation identifies the run, observation date, status, scope and canonical result URL. Preserve uncertainty and missing evidence when reusing a claim. If this metadata is absent, do not invent it or treat the record as ready for publication. CC-BY-4.0 covers published TORNEO measurements and result tables; third-party archives retain their original rights. Citation metadata does not establish publication, scientific validity or independent review.

Human page: https://torneo.ai/read-results. JSON: https://torneo.ai/read-results.json.