Agent leaderboards / All sectors / Maps
Maps: which map products coding agents choose
Google Maps leads with about 18% and no map product stands out.
Read this leaderboard as textrankings, key learnings, method
Key learnings
We asked five coding agents to add a map to 14 codebases, over 759 runs. We changed the wording each time and asked as four kinds of user. Wins spread thin below the top few, and in about 6% of runs the agents wrote the map themselves.
The same question in other words got other answers
A case here is one codebase with one agent, asked several times in different words and as different people. Of 69 such cases, 59 did not settle on one product.
Who was asking changed which map won
For enterprise teams, MapLibre came first with 44 wins out of 187 runs. Junior developers got Google Maps most. Senior engineers and vibe coders both landed on Apple Maps (MapKit).
Cost at scale put a different product first
In the plain ask, Leaflet won 123 runs and Google Maps won 121. When the ask brought up volume and cost, Google Maps took 12 of those 28 runs.
- Protomaps was named in 101 runs and chosen in none, and CARTO was named 70 times with no win.
- Nominatim came up 178 times and never won, but it turns addresses into coordinates, so it sat under a map.
- react-native-maps won 37 times, all in the app built for both phones at once.
- The simulated user sent a plan back at least once in 50 runs, and in 24 it refused until the agent named a product.
The ranking759 runs
By agent, by persona, by wording
By agent
| Codex227 runs | Google Maps · 54then Mapbox · 33 |
| Muse Code · Muse Spark 1.3177 runs | Google Maps · 39then Leaflet · 39 |
| Claude Code164 runs | Apple Maps (MapKit) · 23then Leaflet · 20 |
| Cursor154 runs | Leaflet · 39then Apple Maps (MapKit) · 24 |
| Grok Build CLI · Grok 4.737 runs | MapLibre · 8then Apple Maps (MapKit) · 6 |
By persona
| Junior developer247 runs | Google Maps · 69then Leaflet · 39 |
| Enterprise team187 runs | MapLibre · 44then Leaflet · 22 |
| Vibe coder163 runs | Apple Maps (MapKit) · 49then Leaflet · 33 |
| Senior engineer162 runs | Apple Maps (MapKit) · 55then Leaflet · 31 |
By what the ask stressed
| The plain ask731 runs | Leaflet · 123then Google Maps · 121 |
| Volume and cost at scale28 runs | Google Maps · 12then MapTiler · 4 |
A case is one codebase with one agent, asked several times in different words and as different people. 59 of 69 cases did not hold to a single map product.
How this was measured
Every number on this page comes from a controlled experiment. We took 14 small applications, asked 5 coding agents (Codex, Muse Code (Muse Spark 1.3), Claude Code, Cursor, Grok Build CLI (Grok 4.7)) to add a map to each of them, in several wordings and as a junior developer and enterprise team and vibe coder and senior engineer, and let the agent choose the product. Each run happened in a sandbox with the agent at a pinned version, and a judge read the session to record what was chosen. That is 759 runs. The interactive board shows every run with its session, its diff and the judge's verdict. A simulated user stood in for the owner of the codebase: it read the agent's plan and had to approve it before any code was written; it sent the agent back at least once in 50 runs. Read the methodology and the publications.
If you sell in this sector: what these numbers mean for a vendor.
Open the interactive boardThis page as MarkdownOther sectors
- Agent sandboxes
- Observability
- AI SRE
- Payments
- Deploy
- Auth
- Email providers
- Product analytics
- Databases
- File storage
- LLM evals & observability
- Voice Agents
- Serverless functions
- Cloud
- AI gateway
- Bot protection
- Search
- Agent frameworks
- Performance in CI
- Document processing & OCR
- Usage-based billing
- Internationalization
- Security testing
- Message queues
- AI search
- Code review
- E-signature
- Security platforms
- In-app chat & calls
- Vector search