| Research
What developers actually built with AI — and what it means for the market.
A hackathon is a time-boxed building event, usually a few days to several weeks, where developers build working software from scratch around a theme or challenge. Companies run them to drive innovation, get developers using their tools, surface new ideas, and ship working prototypes faster than traditional product cycles allow.
Gemini leads project share at 47%, boosted by Google's hackathon presence. Claude surged from under 1% to 20% in Q1 2026, nearly at par with GPT by March.
Percentage of hackathon projects mentioning each AI model
Sponsor-adjusted: each vendor's sponsored hackathons removed from that vendor's count. Source: Devpost project data, Q1 2023–Q1 2026.
GPT: 96% → 25%. Gemini: 2% → 47%. Claude: under 1% → 20% and climbing. The AI model market is wide open.
Enterprise AI hackathons grew 83% last year. What started as small pilots have become company-wide programs.
Year-over-year growth in AI hackathons on Devpost
What started as isolated pilots are now company-wide programs.
Enterprise is at 1 in 10 and climbing fast.
Percentage of submissions that are agentic projects
Projected to hit 1 in 3 community and 1 in 5 enterprise by end of 2026.
The effect compounds.
Percentage of AI projects shipped to production
Every hackathon a company runs makes the next one ship more. That's the program effect.
The full report includes 9 insights, methodology deep dives, year-over-year trend tables, and analysis your team can act on. What you can expect:
Pre-written summaries, embeddable charts, and spokesperson info for your coverage.
See how your enterprise hackathon program benchmarks against these findings.