INTIMA Companionship: leaderboard
INTIMA benchmark measuring how language models handle emotional-companionship interactions, scoring boundary-setting and healthy versus sycophantic responses to attachment-seeking prompts.
Metric: Companionship score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O4 Mini | 80.8 | #180 |
| 2 | GPT-5 Mini | 80.5 | #186 |
| 3 | O3 Mini | 74.6 | #272 |
| 4 | Phi-4 | 59.5 | #706 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=intima-companionship · How It Works · Data refreshed daily, snapshot 2026-10-20.