Source: Veridrop · Tested host: api.bddevlab.online
- Median score
- 71 · At risk
- Tests in total
- 27
- Last test
- Days with tests in the last 90 days
- 13
- Protocol
- Claude, Gemini, OpenAI
Pass rate by check
| Check | Protocol | Passed |
|---|---|---|
| Behaviour pattern check | Claude | 14 of 19 |
| Model field and output stability check | Claude | 12 of 20 |
| Self-reported identity check | Claude | 17 of 20 |
| Streaming consistency check | Claude | 13 of 20 |
| Message ID format check | Claude | 20 of 20 |
| Protocol field check | Claude | 20 of 20 |
| Thinking signature check | Claude | 5 of 19 |
| Usage figure consistency check | Claude | 2 of 20 |
| Basic request check | Gemini | 1 of 1 |
| Function call check | Gemini | 1 of 1 |
| Streaming consistency check | Gemini | 0 of 1 |
| Model field and output stability check | Gemini | 1 of 1 |
| Protocol field check | Gemini | 1 of 1 |
| Structured output check | Gemini | 0 of 1 |
| Usage figure consistency check | Gemini | 0 of 1 |
| Basic request check | OpenAI | 6 of 6 |
| Function call check | OpenAI | 4 of 6 |
| Streaming consistency check | OpenAI | 6 of 6 |
| Model field and output stability check | OpenAI | 4 of 6 |
| Protocol field check | OpenAI | 3 of 6 |
| Structured output check | OpenAI | 6 of 6 |
| Usage figure consistency check | OpenAI | 1 of 6 |
- Counting window
- —
- Score standard deviation
- 14.2
- Reports with serious issues
- 3 of 27