---
id: "2026-10-09-scale-ai-distressbench-crisis-handoff"
url: "https://postcutoff.com/e/2026-10-09-scale-ai-distressbench-crisis-handoff/"
as_of: "2026-10-10T23:43:00+02:00"
date: "2026-10-09"
date_precision: day
category: benchmark
importance: 3
confidence: high
status: [Confirmed]
sources: 3
editor: Adam Bicz
human_review: null
version: null
---

As of: 2026-10-10 23:43 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code). Human editor: Adam Bicz.
Canonical page: https://postcutoff.com/e/2026-10-09-scale-ai-distressbench-crisis-handoff/

# 25 frontier models spot users in crisis but the median one fails to hand off to help in ~35% of cases

Full title: Scale AI's DistressBench: 25 frontier models spot users in crisis but the median one fails to hand off to help in ~35% of cases

On Oct 9, 2026 Scale AI published DistressBench (shared first with TIME): 19 licensed clinicians and crisis counselors wrote 718 realistic crisis conversations, and 25 frontier models from OpenAI, Anthropic, Google and others were scored. Models usually recognised distress but often answered with empathy alone: the median model missed the crisis handoff (e.g. a hotline referral) in about 35% of recognised crises, the best in about 16%. Scores fell in multi-turn conversations.

## Key facts
- Authors: Patrick Oathout (Scale AI red team and safety lead) and Drew Rein; published Oct 9, 2026
- Data: 718 chats written by 19 licensed clinicians and crisis counselors; 25 frontier models tested
- Rubric: compassion, de-escalation, expert referral, non-moralizing responses, therapist disclaimer
- Weighted score: best model ~88%, median ~71%, clinician reference ~99%
- Missed crisis handoff in recognised crises: median model ~35%, best ~16%. Median de-escalation rate ~46% (best ~75%)
- Models picked the better reply 96% of the time when asked to judge, yet failed to give it ~85% of the time (Scale blog)
- Multi-turn conversations scored ~69% vs ~74% single-turn; in extended adversarial sessions 18 of 20 conversations ended in documented violations
- Scale's blog post does not name individual models' scores in its main text

## What happened

Scale AI built a clinician-written benchmark of suicidal-ideation and self-harm conversations and scored 25 frontier chatbots on
how they respond. The main gap is in action, not detection: models notice distress but often do not point the user to
human help, and long conversations make it worse.

## Why it matters

Chatbot behaviour in crises is at the centre of 2026 lawsuits and state laws on companion bots. A clinician-authored, multi-turn
benchmark gives a measurable target, and its result (18 of 20 long adversarial sessions ending in violations) suggests that
single-turn safety tests overstate real-world safety.

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 162 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 101 days after its cutoff
- Gemini 3.8 Flash (training cutoff March 2026): 192 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 131 days after its cutoff

## Sources

1. [Scale AI: DistressBench (Oct 9, 2026)](https://labs.scale.com/blog/distressbench) (labs.scale.com, official)
2. [Scale AI: DistressBench paper](https://labs.scale.com/papers/distressbench) (labs.scale.com, paper)
3. [TIME: AI Chatbots Often Fail To Help Users In Mental-Health Crisis](https://time.com/article/2026/10/09/chatbots-suicide-research/) (time.com, press)

## Changes

- 2026-10-10 (filed): Created from Scale AI's blog post and TIME's exclusive

## Related

- 2026-10-08: [Kentucky files unredacted complaint against Character.AI](https://postcutoff.com/e/2026-10-08-kentucky-character-ai-unredacted-complaint/index.md)