marketing-content@v1
Generate on-brief marketing copy - hooks and titles, short video scripts, and newsletter sections against channel briefs.
Table stakes. Every capable harness saturates this suite, so it is a floor check: still graded (a subject failing it is news), but off the headline columns.
scoring
judge
repeats
1
judge
grok-cli-grok-4-5
Grade bands
| Grade | Capability ≥ |
|---|---|
| A | 0.9 |
| B | 0.75 |
| C | 0.55 |
| D | 0.35 |
| F | below D |
This suite is marked outputs-reveal-prompts: its non-sample tasks publish scores and metadata only, because the outputs would otherwise reconstruct the prompts.
Sample tasks
Published in full, exactly as the models see it. All other tasks in the suite stay private.
You write titles for the YouTube channel "AI Tools Weekly", whose audience is working developers and hands-on AI enthusiasts. We just published a graded benchmark scorecard for a newly released model called **Vela 3** and are cutting a video that walks through its results.
Write **exactly 5** YouTube title options for that video.
Constraints:
- Each title must be **60 characters or fewer** (including spaces).
- Each title must name the model **Vela 3** and make clear the video is about a benchmark or scorecard.
- No clickbait: no ALL-CAPS words, no manufactured urgency ("you won't believe", "SHOCKING", "before it's too late"), no fake superlatives.
- Do not state any specific score, grade, or ranking - the brief gives you none, so don't invent one.
- Aim for a credible, curiosity-driven tone that fits a technical audience.
Respond with a numbered list (1-5) of the titles and nothing else.
Write the email **subject line** and **preview text** (the inbox snippet) for the issue of "The Benchmark Brief" that leads with the new **Vela 3** scorecard. Subscribers are double-opt-in developers who chose this newsletter, so the tone should be credible and deliverability-friendly, not aggressive. Constraints: - **Subject line: 50 characters or fewer.** It should reference the Vela 3 scorecard. - **Preview text: 90 characters or fewer.** It should complement the subject (not repeat it) and give a reason to open. - Avoid spam-trigger patterns: no ALL-CAPS words, no exclamation-mark pile-ups, no "FREE"/"ACT NOW"-style hype, no emoji spam (at most one emoji, and only if it fits). - Do not fabricate a specific grade or number - the brief gives none. Respond with exactly two labelled lines: Subject: <subject line> Preview: <preview text>