Research
Read what SHIP measured: benchmarks and studies of coding agents, with the data published beside them.
Latest
|Research
Codex vs Claude Code in 2026: 120 missions on the same 20 tickets
On 20 tickets in one codebase, codex resolved 55 of 60 for $3.63 per resolved ticket and claude-code 58 of 60 for $4.69. Each ran three models.
|Research
SWE-in-a-team: a coding benchmark for software factories
Our SWE-in-a-team benchmark graded thirteen coding agents and models in an SDLC loop. An Open-Weight builder resolved every ticket at half the price.