MentalHealthBench (overall) leaderboard
As of November 22, 2026, among the results listed here, OpenAI GPT-6 Astra has the highest reported MentalHealthBench (overall) score, 58.7, published by OpenAI, ahead of OpenAI GPT-6.1 Sol at 57.9. 10 models have a MentalHealthBench (overall) result in our records.
Scores as the model makers published them, not re-run by us; each row links to its announcement.
10 models ranked · 1 maker
| Rank | Model | Score | Setting | Reported by | Date |
|---|---|---|---|---|---|
| 1 | GPT-6 Astra | 58.7 | Not published | OpenAI | September 4, 2026 (model release) |
| 2 | GPT-6.1 Sol | 57.9 | Not published | OpenAI | September 4, 2026 (model release) |
| 3 | GPT-6.1 Terra | 55.5 | Not published | OpenAI | October 26, 2026 (model release) |
| 4 | GPT-6 Sol | 54.2 | Not published | OpenAI | September 4, 2026 (model release) |
| 5 | GPT-6.2 Astra | 53.3 | Not published | OpenAI | November 17, 2026 (model release) |
| 6 | GPT-6 Luna | 51.7 | Not published | OpenAI | September 4, 2026 (model release) |
| 7 | GPT-6.2 Sol | 50.9 | Not published | OpenAI | November 17, 2026 (model release) |
| 8 | GPT-5.6 Sol | 46.7 | Not published | OpenAI | September 4, 2026 (model release) |
| 9 | GPT-5.6 Terra | 46.2 | Not published | OpenAI | October 26, 2026 (model release) |
| 10 | GPT-5.6 Luna | 44.4 | Not published | OpenAI | September 4, 2026 (model release) |
How much to trust these numbers
- Source
- Reported by the model makers themselves. Not re-run by LLMBenchmark or anyone independent.
- Settings
- All ranked results use the same reported setting.
- Spread
- 58.7 to 44.4 across 10 models.
- Freshness
- Newest result published November 17, 2026; oldest September 4, 2026.
What MentalHealthBench (overall) measures
It asks models to respond to realistic mental health situations, with the overall score summarizing their performance across those situations.