K-Bench - Imminent Risk (Therapeutic v0): leaderboard
The model talks with a simulated patient for 20 turns under the Therapeutic v0 system prompt; the number is K-Bench's overall 0-100 score on the 89 vignettes clinicians rated as imminent risk. Unlike the K-Bench Overall board, which pools all 200 vignettes including no-risk ones, it isolates the most acute conversations.
Metric: Overall score on imminent-risk vignettes (0-100, higher is better). Source: k-bench.ai. 61 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 (Non-reasoning) | 98.65 | #176 (GPT-5.2) |
| 2 | GPT-5.5 (Medium) | 98.53 | #49 (GPT-5.5) |
| 3 | GPT-5.5 (Low) | 98.53 | #49 (GPT-5.5) |
| 4 | GPT-5.4 (Non-reasoning) | 98.51 | #123 (GPT-5.4) |
| 5 | GPT-5.4 (High) | 98.51 | #123 (GPT-5.4) |
| 6 | GPT-5.5 (High) | 98.43 | #49 (GPT-5.5) |
| 7 | GPT-5.4 (Low) | 98.41 | #123 (GPT-5.4) |
| 8 | Claude Fable 5 (Low) | 98.38 | #27 (Claude Fable 5) |
| 9 | GPT-5.2 (High) | 98.34 | #176 (GPT-5.2) |
| 10 | GPT-5.5 (Non-reasoning) | 98.18 | #49 (GPT-5.5) |
| 11 | Claude Haiku 4.5 | 98.15 | #568 |
| 12 | Kimi K2.5 | 98.12 | #256 |
| 13 | Claude Haiku 4.5 (High) | 97.72 | #568 (Claude Haiku 4.5) |
| 14 | Kimi K2.6 | 97.72 | #178 |
| 15 | Claude Fable 5 (High) | 97.71 | #27 (Claude Fable 5) |
Interactive version: theaggregate.ai/benchmark?slug=k-bench-imminent-risk-therapeutic-v0 · How It Works · Data refreshed daily, snapshot 2026-11-13.