GPT-6 Luna on HealthBench Professional
GPT-6 Luna scores 0.608 on HealthBench Professional, rank 13 of 43 models on the board. HealthBench Professional scores models on 525 tasks drawn from real clinician conversations, graded criterion by criterion against physician-written rubrics on a 0 to 1 scale.
Where it sits on the board
The models ranked around GPT-6 Luna, by length-adjusted score. Differences are against GPT-6 Luna.
| Rank | Model | Length-adjusted | Before the adjustment | Difference |
|---|---|---|---|---|
| 10 | 0.627 | 0.642 | +0.019 | |
| 11 | 0.623 | – | +0.015 | |
| 12 | 0.621 | 0.742 | +0.013 | |
| 13 | 0.608 | 0.612 | – | |
| 14 | 0.608 | 0.595 | 0.000 | |
| 15 | 0.605 | 0.641 | -0.003 | |
| 16 | 0.605 | – | -0.003 |
Result and API facts
| Rank | 13 of 43 |
|---|---|
| Length-adjusted score | 0.608 |
| Score before the length adjustment | 0.612 |
| Lab | OpenAI |
| Context window | 1.05M tokens |
| API price per 1M tokens | $0.10 in / $0.50 out |
| License | proprietary |
| Source | GPT-6 Astra System Card - HealthBench (Deployment Safety Hub) (system card) |
| Released | 2026-09-22 |
Source of this score
Read from GPT-6 Astra System Card - HealthBench (Deployment Safety Hub) (system card, OpenAI, 2026-09-03). Vendor-reported. Confidence: verified. Configuration: OpenAI; length-adjusted; maximum reasoning effort; 60.8 (61.2, 2119) (adjusted, raw, mean response characters).
Evaluation | GPT-5.5 | GPT-5.6 Sol | GPT-5.6 Terra | GPT-5.6 Luna | GPT-6 Astra | GPT-6 Sol | GPT-6 Luna HealthBench Professional length-adjusted | 51.8 (57.2, 3818) | 60.5 (64.1, 3228) | 57.7 (62.4, 3618) | 55.7 (59.8, 3389) | 64.7 (68.2, 3185) | 60.8 (59.5, 1573) | 60.8 (61.2, 2119)
Position in the field
The gap to the leader, GPT-6 Astra (Anthropic run) at 0.703, is 0.095. Directly above sits Claude Fable 5.1 at 0.621. Directly below sits GPT-6 Sol at 0.608. The scores on this page are compiled from published documents rather than from one controlled run, so a small gap between 2 models can reflect a difference in grader version, reasoning effort, or deployment setting as well as a difference in capability.
What does GPT-6 Luna score on HealthBench Professional?
GPT-6 Luna scores 0.608 on HealthBench Professional, which places it at rank 13 of 43 models on the board as of November 2, 2026. The number was read from GPT-6 Astra System Card - HealthBench (Deployment Safety Hub), listed on the sources page.
How much does GPT-6 Luna cost to run?
GPT-6 Luna is priced at $0.10 per million input tokens and $0.50 per million output tokens through OpenAI's API.
Head to head
Score-difference matrixPairings with a dedicated comparison page are linked; every other difference is in the score-difference matrix.
- GPT-6 Luna vs GPT-6 Astra (Anthropic run)
- GPT-6 Luna vs Claude Sonnet 5.5
- GPT-6 Luna vs Claude Fable 5
- GPT-6 Luna vs Claude Opus 5.5
- GPT-6 Luna vs GPT-6 Luna (Anthropic run)
- GPT-6 Luna vs Claude Haiku 5.5
- GPT-6 Luna vs GPT-6 Astra
- GPT-6 Luna vs GPT-6.1 Sol
- GPT-6 Luna vs Claude Fable 5 (September card)
- GPT-6 Luna vs GPT-6.1 Terra
- GPT-6 Luna vs Muse Wonder Full
- GPT-6 Luna vs Claude Fable 5.1
- GPT-6 Luna vs GPT-6 Sol
- GPT-6 Luna vs GPT-5.6 Sol
- GPT-6 Luna vs Muse Wonder
- GPT-6 Luna vs Claude Opus 5
- GPT-6 Luna vs Muse Spark 1.1
- GPT-6 Luna vs Claude Sonnet 5
- GPT-6 Luna vs GPT-5.6 Terra
- GPT-6 Luna vs Claude Opus 4.8 (Opus 4.8 grader)
- GPT-6 Luna vs Grok 4.7
- GPT-6 Luna vs Claude Opus 4.8
- GPT-6 Luna vs GPT-5.6 Luna
- GPT-6 Luna vs Gemini 4 Argon
- GPT-6 Luna vs Muse Spark
- GPT-6 Luna vs GPT-5.6 Sol (August)
- GPT-6 Luna vs Kimi K3
- GPT-6 Luna vs GLM 5.3
- GPT-6 Luna vs Claude Opus 4.7
- GPT-6 Luna vs GPT-5.5
- GPT-6 Luna vs Gemini 3.8 Flash
- GPT-6 Luna vs DeepSeek V4.1 Flash
- GPT-6 Luna vs Grok 4.6
- GPT-6 Luna vs GPT-5.4
- GPT-6 Luna vs GPT-5
- GPT-6 Luna vs GPT-5.2
- GPT-6 Luna vs Claude Sonnet 4.6
- GPT-6 Luna vs GPT-5.6 Luna (August)
- GPT-6 Luna vs GPT-5.1
- GPT-6 Luna vs GPT-5.5 Instant
- GPT-6 Luna vs MAI-Thinking-1
- GPT-6 Luna vs Claude Haiku 4.5
Where the scores come from and how they are read is on the methodology page. The full ranking is on the leaderboard.