GPT-5.3 Chat on HealthBench Hard

rank 6 of 9 · updated August 16, 2026

On the 9-model HealthBench Hard board, GPT-5.3 Chat holds rank 6 with a score of 0.259. The API alias of the non-reasoning ChatGPT default from spring 2026, now deprecated in favor of GPT-5.6. 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-5.3 Chat 0.169 off the lead. One place up is GPT OSS 120B at 0.300. One place down is GPT-5.5 Instant at 0.229. 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-5.3 Chat vs Muse Spark
  0.259 vs 0.428 · Muse Spark by 0.169
• GPT-5.3 Chat vs GPT-5.6 Sol
  0.259 vs 0.331 · GPT-5.6 Sol by 0.072
• GPT-5.3 Chat vs GPT-5.6 Terra
  0.259 vs 0.327 · GPT-5.6 Terra by 0.068
• GPT-5.3 Chat vs GPT-5.6 Luna
  0.259 vs 0.320 · GPT-5.6 Luna by 0.061
• GPT-5.3 Chat vs GPT OSS 120B
  0.259 vs 0.300 · GPT OSS 120B by 0.041
• GPT-5.3 Chat vs GPT-5.5 Instant
  0.259 vs 0.229 · GPT-5.3 Chat by 0.030
• GPT-5.3 Chat vs GPT OSS 20B
  0.259 vs 0.108 · GPT-5.3 Chat by 0.151
• GPT-5.3 Chat vs GPT-5
  0.259 vs 0.016 · GPT-5.3 Chat by 0.243

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.