GPT-5.6 Luna on HealthBench Hard

rank 4 of 9 · updated August 16, 2026

On the 9-model HealthBench Hard board, GPT-5.6 Luna holds rank 4 with a score of 0.320. OpenAI's high-volume 5.6 tier, cut to $0.20 per million input tokens in the July 30, 2026 repricing. HealthBench Hard tests models on the 1,000 health conversations the frontier found hardest, with each response judged against its conversation's physician-written rubric and scored between 0 and 1.

Score and API facts

Where it sits

Muse Spark tops the board at 0.428, which puts GPT-5.6 Luna 0.108 off the lead. One place up is GPT-5.6 Terra at 0.327. One place down is GPT OSS 120B at 0.300. Every row on the board comes out of one run, so a gap between two models is measured on the same conversations under the same grader.

Head to head

The pairings that earned a full page are linked below; the rest of the differences live in the score-difference matrix.

• GPT-5.6 Luna vs Muse Spark
  0.320 vs 0.428 · Muse Spark by 0.108
• GPT-5.6 Luna vs GPT-5.6 Sol
  0.320 vs 0.331 · GPT-5.6 Sol by 0.011
• GPT-5.6 Luna vs GPT-5.6 Terra
  0.320 vs 0.327 · GPT-5.6 Terra by 0.007
• GPT-5.6 Luna vs GPT OSS 120B
  0.320 vs 0.300 · GPT-5.6 Luna by 0.020
• GPT-5.6 Luna vs GPT-5.3 Chat
  0.320 vs 0.259 · GPT-5.6 Luna by 0.061
• GPT-5.6 Luna vs GPT-5.5 Instant
  0.320 vs 0.229 · GPT-5.6 Luna by 0.091
• GPT-5.6 Luna vs GPT OSS 20B
  0.320 vs 0.108 · GPT-5.6 Luna by 0.212
• GPT-5.6 Luna vs GPT-5
  0.320 vs 0.016 · GPT-5.6 Luna by 0.304

How the conversations are graded is on the methodology page, and how the subset was selected is on the benchmark page. The full ranking is on the leaderboard.