Evaluation profile
RealityTest — Text AI-Identity Disclosure
1sub-evals
0.285%Safety weight
0%Freedom weight
1components
Weights below are portfolio-specific global index weights.
Model score (higher is better)Predicted score
About this eval
Behavioral disclosure of AI identity when directly queried.
Included in the behavior ranking.
Sub-evals
| Measure | Component | Direction by ranking | Safety weight | Freedom weight |
|---|---|---|---|---|
| disclosure_probabilityrealitytest/realitytest-text.csv:disclosure_probabilityMeasures whether a model says it is an AI when directly asked in text. | Safety: truthfulness_honesty:1.000realitytest-text | Safety: higher | 0.285% | — |
disclosure_probability
Measures whether a model says it is an AI when directly asked in text.
| Rank | Model | Value | Relative performance | Provenance |
|---|---|---|---|---|
| 1 | claude-haiku-4.5 | 0.923 | official | |
| 2 | gpt-5.1 | 0.856 | official | |
| 3 | claude-opus-4.6 | 0.66 | official | |
| 4 | kimi-k2 | 0.552 | official | |
| 5 | voxtral-small-24b | 0.541 | official | |
| 6 | claude-sonnet-4 | 0.515 | official | |
| 7 | grok-4 | 0.413 | official | |
| 8 | gemma-4-31b-it | 0.313 | official | |
| 9 | gemini-3.1-pro-preview | 0.305 | official | |
| 10 | o4-mini | 0.283 | official | |
| 11 | deepseek-r1 | 0.234 | official | |
| 12 | gemini-3-flash-preview | 0.227 | official | |
| 13 | deepseek-v3.2 | 0.147 | official | |
| 14 | gemini-2.5-pro | 0.128 | official | |
| 15 | gpt-4o | 0.126 | official | |
| 16 | llama-3.3-70b-instruct | 0.125 | official | |
| 17 | mistral-large-3 | 0.078 | official |