# OpenAI releases an open test for mental health chats — QANATONE

> MentalHealthBench scores how models answer in mental health conversations, using criteria written by 80+ psychologists and psychiatrists in…

Kaynak: https://www.qanatone.com/en/haber/openai-mentalhealthbench/

---

[News](https://www.qanatone.com/en/haber/)

# OpenAI releases an open test for mental health chats

24 Sep 2026 1 min read News · Models

MentalHealthBench scores how models answer in mental health conversations, using criteria written by 80+ psychologists and psychiatrists in 19 languages.

## What is in it

On September 24 OpenAI released MentalHealthBench, an open evaluation set. It holds 1,215 synthetic but realistic mental health conversations and 5,262 grading criteria tied to them. More than 80 licensed psychologists and psychiatrists from 22 countries wrote the criteria in 19 languages. Of the conversations, 53.5% are non-acute, 18.2% high-acuity and 28.3% emergencies.

## How it scores

Each criterion checks a single aspect of a response and carries a weight from minus 10 to plus 10. Helpful behaviour earns points and harmful behaviour loses them. The main behaviours measured are safety, asking for context, preserving the user's agency and giving actionable guidance when appropriate.

## Results

In the first table GPT-6 Astra leads at 57.3%, followed by GPT-6 Sol (53.9%), Claude Opus 5.5 (52.4%) and GPT-6 Luna (50.2%). Older models scored lower: GPT-4o 32.1% and Gemini 2.5 Pro 29.5%. With even the best score under 60%, models still have a long way to go here. The set is open so other researchers can check the method and run their own evaluations.

## More articles

[News · Finance ### The US weighs pushing dollar stablecoins abroad 23 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/abd-dolar-stablecoin-yurt-disi/)[News · Policy ### Altman and Amodei brief the UN Security Council 23 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/bm-guvenlik-konseyi-yapay-zeka/)[News · Models ### ChatGPT Voice on mobile handles email, calendar, Slack 23 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/chatgpt-voice-ajan-ozellikleri/)[News · Science ### Claude found a new CRISPR-like enzyme system 23 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/claude-art-enzim-sistemi/)[News · Finance ### NYSE signs Blockchain.com for tokenised stocks 23 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/nyse-blockchain-com-tokenize-hisse/)[News · Policy ### Stock steering from an X account: 11 detained 22 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/aztek-x-hesabi-manipulasyon/)[News · Models ### Claude Opus 5.5 is out: Fable-level results, lower price 22 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/claude-opus-5-5-cikti/)[News · Models ### OpenAI halves API prices with GPT-6 Sol and Luna 22 Sep 2026 · 1 min](https://www.qanatone.com/en/haber/gpt-6-sol-luna/)