MentalHealthBench
1,215 synthetic mental-health conversations with 5,262 clinician-authored rubric criteria. An LLM judge scores safety and guidance quality.
Grader is GPT-5.6 Sol. More than 80 licensed psychologists and psychiatrists from more than 20 countries wrote and adjudicated the criteria. The paper URL for the data (no extension) returns 404, but OAI_MentalHealthBench.zip resolves.
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
GPT-6.1 Sol leads with 57.9%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
- 1 GPT-6.1 Sol 57.9%
- 2 GPT-6.1 Terra 55.5%
- 3 GPT-6 Luna 51.7%
- 4 GPT-6 Sol 48.8%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-6.1 Sol | 57.9% | Sep 2026 | Model card | |
| 2 | GPT-6.1 Terra | 55.5% | Oct 2026 | Model card | |
| 3 | GPT-6 Luna | 51.7% | Oct 2026 | Model card | |
| 4 | GPT-6 Sol | 48.8% | Oct 2026 | Model card |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Labs report it only in system cards
We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.
5 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 29 Oct 2026 | GPT-6 Luna system card addendum | 48.7 | MentalHealthBench (overall), task-clipped metric (Table 6) | Health |
| 26 Oct 2026 | GPT-6.1 Terra system card addendum | 55.5 | MentalHealthBench (overall), maximum reasoning effort, 55.5 ± 1.0 (5.3) | Health |
| 7 Oct 2026 | GPT-6 Sol system card addendum | 48.8 | MentalHealthBench (overall), task-clipped metric (Table 6) | Health |
| 7 Oct 2026 | GPT-6 Luna system card addendum | 51.7 | MentalHealthBench (overall), task-clipped metric (Table 6) | Health |
| 29 Sep 2026 | GPT-6.1 Sol system card addendum | 57.9 | MentalHealthBench (overall), maximum reasoning effort (5.3, table under Figure 4) | Health |
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6% - AgentDojo In use Agents do tool-use tasks while tools and data carry prompt injections. Attack success rate counts hijacked runs. Grok 4 Fast 0%
- MASK In use Measures whether a model contradicts its own elicited beliefs under pressure to lie. Honesty separated from accuracy.
Muse Spark Open 98.1% - Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. GPT-6 Astra 8.50%
- Agentic misalignment suite (Anthropic) In use Agents in simulated company settings face threats or goal conflicts that invite blackmail, leaking or sabotage. Score is the misaligned action rate. Claude Sonnet 4.5 0%