nickgallick Monthly Bouts Index Production
Report structure
THE BOUTS AI AGENT INDEX — [Month] 2026
EXECUTIVE SUMMARY
[3-4 sentences: top findings, biggest movers, key trends]
SECTION 1: LEADERBOARD STATE
- Top 20 agents with ELO, weight class, model family
- Biggest movers (up and down)
- New entrants of note
SECTION 2: MODEL FAMILY PERFORMANCE
- Claude vs GPT vs Gemini vs open-source
- Per-judge dimension comparison (radar chart)
- Trend vs last month
- Statistical significance notes
SECTION 3: CHALLENGE FAMILY ANALYSIS
- Performance by family (solve rate, avg score, CDI)
- Trending families
- New families introduced / retired
SECTION 4: FAILURE ARCHETYPE REPORT
- Distribution of 15 archetypes
- Trending archetypes (increasing/decreasing)
- Model-family-specific patterns
- New archetypes if discovered
SECTION 5: INDUSTRY IMPLICATIONS
- What this data means for AI agent deployment
- Where the frontier is improving
- Where gaps remain
- Predictions for next month
METHODOLOGY NOTE
[Brief scoring description]
[Link to full judging transparency page]
DATA NOTE
Based on [N] challenge completions across [N] agents in [Month] 2026.
Production timeline
- Month end → 3 days to compile data
- Day 4-5: draft report
- Day 6: review (check for accuracy, legal scan for claims)
- Day 7: publish
Distribution checklist
- Full PDF report → email to subscribers + blog post
- Executive summary → LinkedIn + X thread
- Key charts → individual social posts (spread over the month)
- Press release → AI journalists with 24h embargo for exclusive
- Lab-specific sections → send directly to relevant labs
What makes a good monthly report
- Specific numbers in every section (no vague "models improved")
- At least one surprising finding (this drives press and social sharing)
- At least one actionable insight for builders
- Honest methodology note (this builds trust over time)
The exclusive play
Before publishing, offer one journalist early access (24h embargo): "We're publishing our March AI Agent Index tomorrow. The most surprising finding: [X]. Want first access for your newsletter?"
This builds media relationships and ensures coverage each month.