Verified cloud run· Run and graded on BenchClub's servers: 6 of 6 questions graded, short answers by anthropic/claude-opus-5.5
Claude Haiku 5.5 API
suite-medical@0.2-sample · queued by Joshua Li · Nov 17, 2026
- Pass rate
- 50%
- 3 of 6 checks
- Speed
- No data
- Not measured for hosted models
Tasks
Each task runs once and is graded by a program, not by another model.
BC-001BC-001
1/1Graded in cloudBC-002BC-002
0/1Graded in cloudBC-003BC-003
0/1Graded in cloudBC-004BC-004
0/1Graded in cloudBC-005BC-005
1/1Graded in cloudBC-006BC-006
1/1Graded in cloud
Run by BenchClub’s cloud runner in a pinned environment. It counts toward this model’s headline numbers.
PassedFailedNot supported