GPT-5.6 Luna (August) on HealthBench Professional
rank 19 of 22 · updated September 8, 2026
GPT-5.6 Luna (August) scores 0.441 on HealthBench Professional, rank 19 of 22 models on the board. The small, high-volume tier of the GPT-5.6 family, repriced to $0.20 per million input tokens on July 30, 2026. 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.
Result and API facts
| rank | 19 of 22 |
|---|---|
| score | 0.441 |
| lab | OpenAI |
| context window | 1.1M tokens |
| API price per 1M tokens | $0.20 in / $1.20 out |
| license | proprietary |
| source | GPT-5.6 - August Updates (system card addendum) (system card) |
| released | 2026-07-09 |
Position in the field
The gap to the leader, Claude Fable 5 at 0.660, is 0.219. Directly above sits Claude Sonnet 4.6 at 0.442. Directly below sits GPT-5.1 at 0.396. 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.
Source of this score
Read from GPT-5.6 - August Updates (system card addendum) (system card, OpenAI, 2026-08-06). Vendor-reported. Confidence: verified. Configuration: ChatGPT production (Instant) deployment setting, length-adjusted, GPT-5.6 August Updates PDF column GPT-5.6 Luna (August) (46.8 unadjusted, 2,920 chars).
HealthBench Professional 32.9 (33.8, 2,285) 38.4 (40.7, 2,775) 54.0 (56.6, 2,894) 44.1 (46.8, 2,920)
p. 11, section 5.1 HealthBench, table "Reported as length-adjusted score (unadjusted, mean response length in characters)", column GPT-5.6 Luna (August) · full entry on the sources page
What does GPT-5.6 Luna (August) score on HealthBench Professional?
GPT-5.6 Luna (August) scores 0.441 on HealthBench Professional, which places it at rank 19 of 22 models on the board as of September 8, 2026. The number was read from GPT-5.6 - August Updates (system card addendum), listed on the sources page.
How much does GPT-5.6 Luna (August) cost to run?
GPT-5.6 Luna (August) is priced at $0.20 per million input tokens and $1.20 per million output tokens through OpenAI's API.
Head to head
Pairings with a dedicated comparison page are linked; every other difference is in the score-difference matrix.
- GPT-5.6 Luna (August) vs Claude Fable 50.441 vs 0.660 · Claude Fable 5 by 0.219
- GPT-5.6 Luna (August) vs GPT-6 Astra0.441 vs 0.634 · GPT-6 Astra by 0.193
- GPT-5.6 Luna (August) vs Claude Fable 5.10.441 vs 0.621 · Claude Fable 5.1 by 0.180
- GPT-5.6 Luna (August) vs GPT-5.6 Sol0.441 vs 0.605 · GPT-5.6 Sol by 0.164
- GPT-5.6 Luna (August) vs Claude Opus 50.441 vs 0.598 · Claude Opus 5 by 0.157
- GPT-5.6 Luna (August) vs Muse Spark 1.10.441 vs 0.593 · Muse Spark 1.1 by 0.152
- GPT-5.6 Luna (August) vs Claude Sonnet 50.441 vs 0.578 · Claude Sonnet 5 by 0.137
- GPT-5.6 Luna (August) vs GPT-5.6 Terra0.441 vs 0.577 · GPT-5.6 Terra by 0.136
- GPT-5.6 Luna (August) vs Claude Opus 4.80.441 vs 0.558 · Claude Opus 4.8 by 0.117
- GPT-5.6 Luna (August) vs GPT-5.6 Luna0.441 vs 0.557 · GPT-5.6 Luna by 0.116
- GPT-5.6 Luna (August) vs Muse Spark0.441 vs 0.541 · Muse Spark by 0.100
- GPT-5.6 Luna (August) vs GPT-5.6 Sol (August)0.441 vs 0.540 · GPT-5.6 Sol (August) by 0.099
- GPT-5.6 Luna (August) vs Claude Opus 4.70.441 vs 0.519 · Claude Opus 4.7 by 0.078
- GPT-5.6 Luna (August) vs GPT-5.50.441 vs 0.518 · GPT-5.5 by 0.077
- GPT-5.6 Luna (August) vs GPT-5.40.441 vs 0.481 · GPT-5.4 by 0.040
- GPT-5.6 Luna (August) vs GPT-50.441 vs 0.462 · GPT-5 by 0.021
- GPT-5.6 Luna (August) vs GPT-5.20.441 vs 0.459 · GPT-5.2 by 0.018
- GPT-5.6 Luna (August) vs Claude Sonnet 4.60.441 vs 0.442 · Claude Sonnet 4.6 by 0.001
- GPT-5.6 Luna (August) vs GPT-5.10.441 vs 0.396 · GPT-5.6 Luna (August) by 0.045
- GPT-5.6 Luna (August) vs GPT-5.5 Instant0.441 vs 0.384 · GPT-5.6 Luna (August) by 0.057
- GPT-5.6 Luna (August) vs MAI-Thinking-10.441 vs 0.350 · GPT-5.6 Luna (August) by 0.091
Where the scores come from and how they are read is on the methodology page. The full ranking is on the leaderboard.