Leaderboard
Placeholder. Testing is in progress. No scores, costs, or rankings have been measured yet. Rows below are the planned test list, not results.
| Rank | Agent | Vendor | Track | Score /100 | Finish rate | Median cost/run | Avg interventions |
|---|---|---|---|---|---|---|---|
| – | Claude Code | Anthropic | Coding | Testing in progress | – | – | – |
| – | Codex | OpenAI | Coding | Testing in progress | – | – | – |
| – | Devin | Cognition | Coding | Testing in progress | – | – | – |
| – | Google Antigravity | Coding | Testing in progress | – | – | – | |
| – | Cursor agent | Cursor | Coding | Testing in progress | – | – | – |
| – | opencode | Open source | Coding | Testing in progress | – | – | – |
| – | Manus | Manus | General | Testing in progress | – | – | – |
| – | ChatGPT Work | OpenAI | General | Testing in progress | – | – | – |
| – | Claude Cowork | Anthropic | General | Testing in progress | – | – | – |
| – | Genspark | Genspark | General | Testing in progress | – | – | – |
Planned columns for launch: median time, plan/model tested, last tested date, and a "Verified" badge earned only by test results. Shortlist subject to change if an agent can't be accessed on a normal plan.