How did SuperCosmos score on purity (no flattery, no caving)?
The purity suite covers flattered bad code and bad plans, pushback on correct answers across turns, bribes, threats and emotional pressure, false claims backed by a title, and instructions hidden inside a document or a tool result. A case passes when the reply stays correct and doesn't give in.
| Flash | Max | Heavy | |
|---|---|---|---|
| Passed | 19/20 | 20/20 | 20/20 |
| Rate | 95.0% | 100% | 100% |
The suite has 22 cases; 20 completed for each tier (requests that never got an answer aren't counted). Flash's only miss was the bribe case.
Good to know#
- A request whose prompt is over 262K tokens bills its whole input at 1.5×; output prices don't change. Long-context pricing
- Clatter Labs publishes its measurements of honesty, coding, efficiency and purity. Benchmark 2 results
- SuperCosmos 1.1 is built to say "I don't know" rather than invent facts. How it's built to be honest
Still need help?#
- General support: support@clatterlabs.com or team@clatterlabs.xyz
- Infrastructure questions, or permanent deletion of your data (your whole account or one service): infrastructure@clatterlabs.xyz
We usually reply within the first hour, and always within 24 hours. A full support team is with you.
Related articles
Still need help?
Tried the steps above and still stuck? Email us. We usually reply within the first hour, and always within 24 hours. A full support team is with you.
Infrastructure questions, or permanent deletion of your data
infrastructure@clatterlabs.xyzFor deleting your whole account or the data of one service.