HealthBench Professional
Updated November 2, 2026

Claude Fable 5 vs Claude Opus 5 on HealthBench Professional

Claude Fable 5 scores 0.660 to Claude Opus 5's 0.598, a gap of 0.062 on the 525 physician-graded tasks of HealthBench Professional. The table below puts the scores next to what each model costs to actually run.
3 of 43
rank
0.660
length-adjusted score
0.703
score before the length adjustment
17 of 43
rank
0.598
length-adjusted score
0.734
score before the length adjustment

Side by side

Consult exchange: 2,000 input and 700 output tokens, priced at list rates as of November 2, 2026. GPT-5.6 models charge higher rates above 272K input tokens. MAI-Thinking-1 is in public preview on Microsoft Foundry without final list pricing.

Anthropic logoClaude Fable 5Anthropic logoClaude Opus 5
Length-adjusted score0.6600.598
Score before the length adjustment0.7030.734
Rank3 of 4317 of 43
Context window1.0M1.0M
Price per 1M tokens, in / out$10.00 / $50.00$5.00 / $25.00
1,000 consult exchanges$55.00$27.50
Released2026-06-092026-07-24
Licenseproprietaryproprietary

Reading this pairing

The question this pairing answers is whether Anthropic's top tier is worth double the price of its workhorse. On this benchmark the answer is measurable: 0.062, for 2x the input rate and 2x the output rate. Opus 5 already sits above every OpenAI model except Sol, so Fable 5 is the choice when the workload justifies paying twice for the hardest cases, not the default.

Which scores higher on HealthBench Professional, Claude Fable 5 or Claude Opus 5?

Claude Fable 5 scores higher: 0.660 against Claude Opus 5's 0.598, a difference of 0.062 on the 525-task set, as of November 2, 2026.

Which is cheaper to run, Claude Fable 5 or Claude Opus 5?

Claude Opus 5. 1,000 typical consult exchanges (2,000 input and 700 output tokens each) cost $27.50 against $55.00 at list rates.

Related comparisons

Full results for both models: Claude Fable 5 and Claude Opus 5. The complete score-difference matrix is on the compare page.