HealthBench Professional
Updated November 2, 2026

Sources

22
documents
43 of 43
rows documented

This page lists the document behind each score on the leaderboard. Arcophos does not run HealthBench Professional; the numbers are copied from vendor system cards, official update reports, model cards, launch announcements, and named leaderboards, exactly as those documents print them. Entries are grouped by document. Each one gives the model, the score as printed, the excerpt the number was read from, where it sits in the document, who reported it, and a confidence label. Rows with no located document are listed at the end rather than left out. How the labels are assigned and why 2 documents can disagree is on the methodology page.

The same source fields are included in the JSON and CSV downloads on the data page.

Addendum to GPT-6 Astra System Card: GPT-6.1 Sol

system card, first party. OpenAI. deploymentsafety.openai.com. Published September 29, 2026. Retrieved October 6, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-6.1 Sol64.20.672
primary sourcevendor-reportedverified

OpenAI; length-adjusted; 64.2 (67.2, 3038) (adjusted, raw, mean response characters); addendum Table 7 refers to the GPT-6 Astra card for method

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 | GPT-6.1 SOL 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) | 64.2 (67.2, 3038)
Section 5.1, Table 7, HealthBench Professional length-adjusted, GPT-6.1 Sol column

Addendum to GPT-6 Astra System Card: GPT-6.1 Terra

system card, first party. OpenAI. deploymentsafety.openai.com. Published October 26, 2026. Retrieved November 2, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-6.1 Terra62.70.642
primary sourcevendor-reportedverified

OpenAI; length-adjusted; 62.7 (64.2, 2531) (adjusted, raw, mean response characters); addendum Table 5 refers to the GPT-6 Astra card for method

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 GPT-6.1-SOL GPT-6.1-TERRA 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) 64.2 (67.2, 3038) 62.7 (64.2, 2531)
Section 5.1, Table 5, HealthBench Professional length-adjusted, GPT-6.1 Terra column

Claude Fable 5 and Claude Mythos 5 System Card

system card, first party. Anthropic. anthropic.com. Published June 9, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Claude Fable 566.00.703
primary sourcevendor-reportedverified

length-adjusted, Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt (raw 70.3%). Measured as Claude Mythos 5; Anthropic itself prints 66.0 in the Fable 5 column of the Opus 5 card with footnote Mythos 5.

HealthBench Professional 66.0 - 64.7 56.9 51.8 -
p. 252, Table 8.1.A, row HealthBench Professional, column Mythos 5 (Fable 5 column '-'); Figure 8.18.2.A p. 298 bar label 66.0% on Claude Mythos 5
Claude Opus 4.856.9–
conflicting
HealthBench Professional 66.0 - 64.7 56.9 51.8 -
p. 252, Table 8.1.A, column Opus 4.8

Claude Fable 5.1 and Claude Mythos 5.1 System Card

system card, first party. Anthropic. anthropic.com.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Claude Fable 563.3%–
conflicting
HealthBench Professional 62.1% 63.3% 59.8% –
p. 167, Table 8.1.A, column 'Claude Fable 5/Mythos 5'
Claude Fable 5 (September card)63.3%0.689
primary sourcevendor-reportedverified

Measured as Claude Fable 5; Anthropic September card; length-adjusted; adaptive max effort; Opus 4.8 grader; five trials; no tools or custom system prompt; raw 68.9%

HealthBench Professional | Length-adjusted score | Claude Fable 5 | 63.3%
p. 199, Figure 8.17.2.A, Claude Fable 5 length-adjusted bar (visually read printed labels).
Claude Fable 5.162.1%0.742
primary sourcevendor-reportedverified

length-adjusted (method published in the HealthBench Professional paper); Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over five trials, no tools or customized system prompt; Fable 5.1 run with safety classifiers active and a refusal-fallback to Claude Opus 5 (raw 74.2%).

On HealthBench Professional, Claude Fable 5.1 achieved a raw score of 74.2%, ahead of Claude Opus 5 at 73.4%, Fable 5 at 68.9%, and Claude Sonnet 5 at 62.4%. After length adjustment, which penalizes verbose model responses, Fable 5.1 achieved a score of 62.1%.
p. 199, sec. 8.17.2 (same figure printed as 62.1% in Table 8.1.A, p. 167)
Claude Opus 559.8%–
corroborating
HealthBench Professional 62.1% 63.3% 59.8% –
p. 167, Table 8.1.A, column 'Claude Opus 5'; raw 73.4% also reprinted on p. 199, sec. 8.17.2

Claude Haiku 5.5 System Card

system card, first party. Anthropic. anthropic.com. Published October 7, 2026. Retrieved October 12, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-6 Luna (Anthropic run)65.5%not printed in the source
primary sourceindependent runverified

Run by Anthropic through OpenAI's public API; max effort; Opus 4.8 grader; length-adjusted; not OpenAI's published figure

HealthBench Professional panel: 65.5% (GPT-6 Luna (OpenAI)); caption: GPT-6 Luna's scores come from our own runs through OpenAI's public API and are not OpenAI's published figures.
p. 132, Figure 8.11.1.A, HealthBench Professional panel, GPT-6 Luna (OpenAI) diamond label (visually read printed label); Section 8.11.1 text on pp. 131-132
Claude Haiku 5.564.8%0.710
primary sourcevendor-reportedverified

Anthropic; max effort; Opus 4.8 grader; safety classifiers enabled; length-adjusted; raw 71.0%; five runs (October 1 and 2); thirteen repeat max runs on October 4 averaged 64.2%

Max effort scored highest, at 64.8% (71.0% raw); thirteen repeat runs at max on October 4, the day of the lower levels, averaged 64.2%.
p. 133, Section 8.11.3; same value in p. 111, Table 8.1.A, row HealthBench Professional (Length-adjusted), column Haiku 5.5 (printed 64.8) and p. 132, Figure 8.11.1.A
Claude Haiku 4.532.2not printed in the source
primary sourcevendor-reportedverified

Anthropic; extended thinking (Haiku 4.5 has no effort setting); Opus 4.8 grader; safety classifiers enabled; length-adjusted; measured for the Haiku 5.5 card

Evaluation | Haiku 5.5 | Haiku 4.5 | Sonnet 5.5 | GPT-6 Luna HealthBench Professional (Length-adjusted) | 64.8 | 32.2 | 69.2 | –
p. 111, Table 8.1.A, row HealthBench Professional (Length-adjusted), column Haiku 4.5

Claude Opus 5.5 System Card

system card, first party. Anthropic. anthropic.com. Published September 22, 2026. Retrieved November 1, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Claude Opus 5.565.6%0.771
primary sourcevendor-reportedverified

Anthropic; adaptive max effort; Opus 4.8 grader; five trials; no tools or custom system prompt; safety classifiers and refusal fallback to Opus 5; length-adjusted; raw 77.1%

8.15.2 HealthBench Professional results length adjustment, which penalizes verbose model responses, Claude Opus 5.5 achieved a score of 65.6%.
pp. 213–214, Section 8.15.2, paragraph and Figure 8.15.2.A.

Claude Sonnet 5.5 System Card

system card, first party. Anthropic. anthropic.com. Published September 28, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-6 Astra (Anthropic run)70.3%0.740
primary sourceindependent runverified

Anthropic public-API reproduction; max effort; no system prompt; Opus 4.8 grader; length-adjusted; raw 74.0%

On HealthBench Professional at max effort, length adjustment changes the ranking. After: GPT-6 Astra (70.3%) > Claude Sonnet 5.5 (69.2%) > Claude Opus 5.5 (65.6%) > Claude Fable 5.1 (62.1%).
p. 138, paragraph above Figure 8.15.B, HealthBench Professional after length adjustment.
Claude Sonnet 5.569.2%0.771
primary sourcevendor-reportedverified

Anthropic; max effort; Opus 4.8 grader; safety classifiers enabled; length-adjusted; raw 77.1%

On HealthBench Professional at max effort, length adjustment changes the ranking. After: GPT-6 Astra (70.3%) > Claude Sonnet 5.5 (69.2%) > Claude Opus 5.5 (65.6%) > Claude Fable 5.1 (62.1%).
p. 138, Section 8.15.2 and paragraph above Figure 8.15.B.

GPT-5.5 Instant System Card

system card, first party. OpenAI. deploymentsafety.openai.com. Published May 5, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-5.5 Instant38.40.407
primary sourcevendor-reportedverified

length-adjusted (40.7 unadjusted, 2,775 mean response chars); GPT-5.5 Instant system card Table 5, column GPT-5.5 INSTANT.

HealthBench Professional 37.6 (40.4, 2,973) 35.7 (38.3, 2,872) 32.9 (33.8, 2,285) 38.4 (40.7, 2,775)
Section 4.1 HealthBench, Table 5 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.5 INSTANT

GPT-5.6 - August Updates (system card addendum)

system card, first party. OpenAI. cdn.openai.com. Published August 6, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-5.6 Sol (August)54.00.566
primary sourcevendor-reportedverified

ChatGPT production (Instant) deployment setting, length-adjusted, GPT-5.6 August Updates PDF column GPT-5.6 Sol (August) (56.6 unadjusted, 2,894 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 Sol (August)
GPT-5.6 Luna (August)44.10.468
primary sourcevendor-reportedverified

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)
GPT-5.5 Instant38.4–
corroborating
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-5.5 Instant

GPT-5.6 Preview System Card

system card, first party. OpenAI. deploymentsafety.openai.com.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-5.6 Sol60.5–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5.6-SOL
GPT-5.6 Terra57.7–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5.6-TERRA
GPT-5.6 Luna55.7–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5.6-LUNA
GPT-5.551.8–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5.5
GPT-5.448.1–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5.4
GPT-546.2–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5
GPT-5.245.9–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5.2
GPT-5.139.6–
corroborating
HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)) (preview card, published 2026-06-26), column GPT-5.1

GPT-5.6 System Card

system card, first party. OpenAI. deploymentsafety.openai.com. Published July 9, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-5.6 Sol60.50.641
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort (64.1 unadjusted, 3,228 mean response chars); GPT-5.6 system card Table 6, column GPT-5.6-SOL. API max-effort setting, not the August ChatGPT production setting measured in row 492.

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.6-SOL
GPT-5.6 Terra57.70.624
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort (62.4 unadjusted, 3,618 mean response chars); GPT-5.6 system card Table 6, column GPT-5.6-TERRA.

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.6-TERRA
GPT-5.6 Luna55.70.598
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort (59.8 unadjusted, 3,389 mean response chars); GPT-5.6 system card Table 6, column GPT-5.6-LUNA. API max-effort setting, not the August ChatGPT production setting measured in row 493.

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.6-LUNA
GPT-5.551.80.572
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.5 (57.2 unadjusted, 3818 chars)

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.5
GPT-5.448.10.519
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.4 (51.9 unadjusted, 3308 chars)

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.4
GPT-546.20.510
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5 (51.0 unadjusted, 3616 chars)

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5
GPT-5.245.90.500
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.2 (50.0 unadjusted, 3400 chars)

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.2
GPT-5.139.60.480
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort, GPT-5.6 system card Table 6 column GPT-5.1 (48.0 unadjusted, 4863 chars)

HealthBench Professional length-adjusted 46.2 (51.0, 3616) 39.6 (48.0, 4863) 45.9 (50.0, 3400) 48.1 (51.9, 3308) 51.8 (57.2, 3818) 60.5 (64.1, 3228) 57.7 (62.4, 3618) 55.7 (59.8, 3389)
Section 5.1 HealthBench, Table 6 (reported as length-adjusted score (unadjusted, mean response length in characters)), column GPT-5.1

GPT-6 Astra System Card

system card, first party. OpenAI. deploymentsafety.openai.com.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-6 Astra63.4–
conflicting
Astra has a length-adjusted HealthBench Professional score of 63.4 (+2.9 relative to GPT-5.6 Sol), HealthBench score of 58.1 (+1.1), HealthBench Hard score of 36.3 (+3.2), and HealthBench Consensus score of 95.8 (+0.3).
p. 19, sec. 6.1 (Table 6 cell, column 'gpt-6 Astra': 63.4 (69.5, 4097))

GPT-6 Astra System Card - HealthBench (Deployment Safety Hub)

system card, first party. OpenAI. deploymentsafety.openai.com. Published September 3, 2026. Retrieved September 8, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-6 Astra64.70.682
primary sourcevendor-reportedverified

length-adjusted, max reasoning effort (68.2 unadjusted, 3,185 mean response chars, as corrected by OpenAI on September 22; first printed as 69.5 and 4,097); GPT-6 Astra system card Table 6, column 'gpt-6 Astra'.

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)
Section 11.4.1, Table 29, HealthBench Professional length-adjusted, GPT-6 Astra column; September 22 correction.
GPT-6 Astra63.4–
mirror
Astra has a length-adjusted HealthBench Professional score of 63.4 (+2.9 relative to GPT-5.6 Sol)
section 6.1, HTML rendering of the same card
GPT-6 Luna60.80.612
primary sourcevendor-reportedverified

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)
Section 11.4.1, Table 29, HealthBench Professional length-adjusted, GPT-6 Luna column; September 22 correction.
GPT-6 Sol60.80.595
primary sourcevendor-reportedverified

OpenAI; length-adjusted; maximum reasoning effort; 60.8 (59.5, 1573) (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)
Section 11.4.1, Table 29, HealthBench Professional length-adjusted, GPT-6 Sol column; September 22 correction.

GPT-6 Astra: A new generation of intelligence

launch post, first party. OpenAI. openai.com.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
GPT-6 Astra63.4%–
corroborating
| HealthBench Professional (length-adjusted) | 63.4% | 60.5% | 58.1% | 60.9% | 56.4% | 52.1% |
Science and Health table

HealthBench Professional Benchmark | Artificial Analysis

independent leaderboard. Artificial Analysis. artificialanalysis.ai. Published November 2, 2026. Retrieved November 2, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Gemini 4 Argon55.1%0.642
primary sourceindependent runpartial

Artificial Analysis independent run, high reasoning; GPT-5.4 (low) judge, 4 repeats per task; length-adjusted (64.2 raw, 5,133 mean response chars)

HealthBench Professional: Length-Adjusted Score | Gemini 4 Argon (high) | 55.1%
Score section, Length-Adjusted Score chart, Gemini 4 Argon (high) bar label; raw score 64.2% on the Raw Score tab
Kimi K353.2%0.569
primary sourceindependent runpartial

Artificial Analysis independent run, max reasoning; GPT-5.4 (low) judge, 4 repeats per task; length-adjusted (56.9 raw, 3,250 mean response chars)

HealthBench Professional: Length-Adjusted Score | Kimi K3 (max) | 53.2%
Score section, Length-Adjusted Score chart, Kimi K3 (max) bar label; raw score 56.9% on the Raw Score tab
GLM 5.352.8%0.579
primary sourceindependent runpartial

Artificial Analysis independent run, max reasoning; GPT-5.4 (low) judge, 4 repeats per task; length-adjusted (57.9 raw, 3,748 mean response chars)

HealthBench Professional: Length-Adjusted Score | GLM-5.3 (max) | 52.8%
Score section, Length-Adjusted Score chart, GLM-5.3 (max) bar label; raw score 57.9% on the Raw Score tab
Gemini 3.8 Flash51.7%0.590
primary sourceindependent runpartial

Artificial Analysis independent run, high reasoning; GPT-5.4 (low) judge, 4 repeats per task; length-adjusted (59.0 raw, 4,494 mean response chars)

HealthBench Professional: Length-Adjusted Score | Gemini 3.8 Flash (high) | 51.7%
Score section, Length-Adjusted Score chart, Gemini 3.8 Flash (high) bar label; raw score 59.0% on the Raw Score tab
DeepSeek V4.1 Flash50.7%0.529
primary sourceindependent runpartial

Artificial Analysis independent run, max reasoning; GPT-5.4 (low) judge, 4 repeats per task; length-adjusted (52.9 raw, 2,760 mean response chars)

HealthBench Professional: Length-Adjusted Score | DeepSeek V4.1 Flash (max) | 50.7%
Score section, Length-Adjusted Score chart, DeepSeek V4.1 Flash (max) bar label; raw score 52.9% on the Raw Score tab

Introducing Grok 4.7

launch post, first party. SpaceXAI. x.ai. Published September 21, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Grok 4.756.7%not printed in the source
primary sourcevendor-reportedverified

SpaceXAI evaluation; xHigh effort; HealthBench Professional

Grok 4.7 xHigh Grok 4.6 High GPT-5.6 Sol Max Fable 5.1 Max Clinical reasoningHealthBench Professional 56.7% 48.5% 60.5% 62.1%
Model Improvements comparison table, Clinical reasoning / HealthBench Professional row; Grok 4.7 column; September 21, 2026.
Grok 4.648.5%not printed in the source
primary sourcevendor-reportedverified

SpaceXAI evaluation; High effort; HealthBench Professional

Grok 4.7 xHigh Grok 4.6 High GPT-5.6 Sol Max Fable 5.1 Max Clinical reasoningHealthBench Professional 56.7% 48.5% 60.5% 62.1%
Model Improvements comparison table, Clinical reasoning / HealthBench Professional row; Grok 4.6 column; September 21, 2026.

MAI-Thinking-1: Building a Hill-Climbing Machine

model card, first party. Microsoft AI. microsoft.ai. Published August 12, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
MAI-Thinking-135not printed in the source
primary sourcevendor-reportedverified

length-adjusted (HealthBench Professional length penalty), standard GPT-5.4 grader and OpenAI rubrics, Microsoft AI run; printed at integer precision.

Model AIR-Bench CyberSec Instruct CyberSec Auto Long Fact Truthful QA HealthBench Prof. MedXpert QA MAI-Thinking-1 88 63 63 98 88 35 43 Sonnet 4.6 88 62 56 98 88 38 49
p. 54, Table 12 'Post-trained model evaluation results on various public benchmarks', Health group, column HealthBench Prof.; protocol Appendix K.6 p. 106: HealthBench Professional introduces a length penalty for the primary metric, to correct for a well-observed correlation between lengthy responses and artificially increased LLM-grader scores. For all reported scores, we use the standard GPT-5.4 grader and rubrics provided by OpenAI.

Muse Spark 1.1 Evaluation Report

model card, first party. Meta. research.meta.ai. Published July 9, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Muse Spark 1.159.3not printed in the source
primary sourcevendor-reportedverified

length-normalized, GPT-5.4 low-reasoning grader, xhigh reasoning via Meta Model API (Muse Spark 1.1 Evaluation Report Figure 44)

Health | HealthBench Professional | 59.3 | 54.1 | 41.6 | 55.8 | 51.8 (Figure 44 image table row; columns Muse Spark 1.1, Muse Spark, Gemini 3.1 Pro (high), Opus 4.8 (max), GPT 5.5 (xhigh))
p. 101, Figure 44 'General capability benchmark results' (image), row HealthBench Professional, column Muse Spark 1.1; protocol p. 104 (printed 103): HealthBench Pro comprises 525 evaluation data points graded by rubrics. We use GPT-5.4 with low reasoning effort as the grader and report the length-normalized rubric score as done in their paper.
Muse Spark54.1not printed in the source
primary sourcevendor-reportedverified

length-normalized, GPT-5.4 low-reasoning grader; Muse Spark (1.0) column in the Muse Spark 1.1 Evaluation Report Figure 44

Health | HealthBench Professional | 59.3 | 54.1 | 41.6 | 55.8 | 51.8 (Figure 44 image table row; columns Muse Spark 1.1, Muse Spark, Gemini 3.1 Pro (high), Opus 4.8 (max), GPT 5.5 (xhigh))
p. 101, Figure 44 (image), row HealthBench Professional, column Muse Spark; protocol p. 104 (printed 103)

Muse Wonder Model Card

model card, first party. Meta. research.meta.ai. Published October 12, 2026. Retrieved November 2, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Muse Wonder Full62.3not printed in the source
primary sourcevendor-reportedverified

Meta internal run; length-normalized weighted rubric score, GPT-5.4 grader, no tools, max reasoning effort (Muse Wonder Model Card p. 3 table, Muse Wonder Full (max) column)

HealthBench Pro (Open-ended health queries) | Muse Wonder Full (max) 62.3 | Muse Wonder (max) 60.5 | Claude Fable 5.1 (max) 62.1 | GPT-6 Astra (max) 65.3 | Claude Opus 5.5 (max) 65.6 Methodology (p. 4): HealthBench Pro evaluates responses to 525 professional-grade health questions. Models answer without tools, and a GPT-5.4 grader scores each response against physician-written rubrics. We run the official tasks and grading in a common internal framework. The primary metric is the length-normalized weighted rubric score.
p. 3, results table (image), HealthBench Pro row, Muse Wonder Full (max) column; methodology p. 4
Muse Wonder60.5not printed in the source
primary sourcevendor-reportedverified

Meta internal run; length-normalized weighted rubric score, GPT-5.4 grader, no tools, max reasoning effort (Muse Wonder Model Card p. 3 table, Muse Wonder (max) column)

HealthBench Pro (Open-ended health queries) | Muse Wonder Full (max) 62.3 | Muse Wonder (max) 60.5 | Claude Fable 5.1 (max) 62.1 | GPT-6 Astra (max) 65.3 | Claude Opus 5.5 (max) 65.6 Methodology (p. 4): HealthBench Pro evaluates responses to 525 professional-grade health questions. Models answer without tools, and a GPT-5.4 grader scores each response against physician-written rubrics. We run the official tasks and grading in a common internal framework. The primary metric is the length-normalized weighted rubric score.
p. 3, results table (image), HealthBench Pro row, Muse Wonder (max) column; methodology p. 4

System Card: Claude Opus 4.8

system card, first party. Anthropic. anthropic.com. Published May 28, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Claude Opus 4.855.8%not printed in the source
primary sourcevendor-reportedverified

length-adjusted, adaptive thinking at max effort, Claude Sonnet 4.6 grader (the grader used by the Opus 4.8 card itself); the other Claude rows on this board use the Claude Opus 4.8 grader, under which Anthropic later prints 57.4 for this model.

Claude Opus 4.8 scores 55.8%, a meaningful improvement over Claude Opus 4.7 at 51.9% and Claude Sonnet 4.6 at 41.7%.
p. 228, section 8.14.1 HealthBench Professional; Figure 8.14.A p. 229
Claude Opus 4.751.9%not printed in the source
primary sourcevendor-reportedverified

length-adjusted; adaptive thinking at max effort; Claude Sonnet 4.6 grader; 5 trials; comparison model in the Opus 4.8 card

Claude Opus 4.8 scores 55.8%, a meaningful improvement over Claude Opus 4.7 at 51.9% and Claude Sonnet 4.6 at 41.7%.
p. 228, section 8.14.1 HealthBench Professional; Figure 8.14.A p. 229
Claude Sonnet 4.641.7%–
conflicting
Claude Opus 4.8 scores 55.8%, a meaningful improvement over Claude Opus 4.7 at 51.9% and Claude Sonnet 4.6 at 41.7%.
p. 228, section 8.14.1

System Card: Claude Opus 5

system card, first party. Anthropic. anthropic.com.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Claude Fable 566.0–
corroborating
HealthBench Professional 59.8 57.4 66.011 60.5
p. 152, Table 8.1.A, column Fable 5 (footnote 11: Mythos 5)
Claude Opus 559.8%0.734
primary sourcevendor-reportedverified

length-adjusted, Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt (raw 73.4%).

Claude Opus 5 achieved a raw score of 73.4%, which is the highest amongst all Claude models, ahead of Claude Mythos 5 at 70.3%, Claude Opus 4.8 at 60.3%, and Claude Sonnet 5 at 62.4%. After length adjustment, which penalizes verbose model responses, Claude Opus 5 achieved a score of 59.8%.
p. 189, section 8.15.2 HealthBench Professional results; also Table 8.1.A p. 152 ('HealthBench Professional 59.8 ...')
Claude Sonnet 557.8%–
corroborating
Figure 8.15.2.A bar labels: Claude Sonnet 5 62.4% raw / 57.8% length-adjusted (prose: 'Claude Sonnet 5 at 62.4%')
p. 189, section 8.15.2, Figure 8.15.2.A
Claude Opus 4.857.4–
conflicting
HealthBench Professional 59.8 57.4 66.011 60.5
p. 152, Table 8.1.A, column Opus 4.8

System Card: Claude Sonnet 5

system card, first party. Anthropic. anthropic.com. Published June 30, 2026. Retrieved September 7, 2026.

ModelLength-adjusted scoreScore before the length adjustmentSource detailsExcerpt and locator
Claude Sonnet 557.80.624
primary sourcevendor-reportedverified

length-adjusted, Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt (raw 62.4%).

HealthBench Professional 57.8 44.2 51.8 -
p. 115, Table 8.1.A, row HealthBench Professional, column Claude Sonnet 5; Figure 8.12.2.A p. 139
Claude Opus 4.8 (Opus 4.8 grader)57.4%not printed in the source
primary sourcevendor-reportedverified

Anthropic June evaluation; length-adjusted; adaptive max effort; Opus 4.8 grader; five trials; no tools or custom system prompt

HealthBench Professional | Length-adjusted score (%) | Claude Opus 4.8 | 57.4%
p. 139, Figure 8.12.2.A, Claude Opus 4.8 bar (visually read printed labels).
Claude Sonnet 4.644.2not printed in the source
primary sourcevendor-reportedverified

length-adjusted; Anthropic protocol: adaptive thinking at max effort, Claude Opus 4.8 grader, averaged over 5 trials, no tools or custom system prompt; comparison column in the Sonnet 5 card. Other Anthropic prints: 44.4% (Fable card Figure 8.18.2.A), 41.7% (Opus 4.8 card, Sonnet 4.6 grader)

HealthBench Professional 57.8 44.2 51.8 -
p. 115, Table 8.1.A, row HealthBench Professional, column Claude Sonnet 4.6