HealthBench Professional

Anthropic logoClaude Haiku 4.5 on HealthBench Professional

Rank 43 of 43, updated November 2, 2026.

Claude Haiku 4.5 scores 0.322 on HealthBench Professional, rank 43 of 43 models on the board. 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.

43
rank of 43
0.322
length-adjusted score
–
score before the length adjustment, not printed in the source
0.381
gap to the leader, GPT-6 Astra (Anthropic run)
200K
context window, tokens
$1.00 / $5.00
input / output price per 1M tokens

Where it sits on the board

The models ranked around Claude Haiku 4.5, by length-adjusted score. Differences are against Claude Haiku 4.5.

RankModelLength-adjustedBefore the adjustmentDifference
37OpenAI logoGPT-5.20.4590.500+0.137
38Anthropic logoClaude Sonnet 4.60.442–+0.120
39OpenAI logoGPT-5.6 Luna (August)0.4410.468+0.119
40OpenAI logoGPT-5.10.3960.480+0.074
41OpenAI logoGPT-5.5 Instant0.3840.407+0.062
42Microsoft logoMAI-Thinking-10.350–+0.028
43Anthropic logoClaude Haiku 4.50.322––

Result and API facts

Rank43 of 43
Length-adjusted score0.322
Score before the length adjustmentnot printed in the source
LabAnthropic
Context window200K tokens
API price per 1M tokens$1.00 in / $5.00 out
Licenseproprietary
SourceClaude Haiku 5.5 System Card (system card)
Released2025-10-15

Source of this score

Read from Claude Haiku 5.5 System Card (system card, Anthropic, 2026-10-07). Vendor-reported. Confidence: verified. Configuration: 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.

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

Position in the field

The gap to the leader, GPT-6 Astra (Anthropic run) at 0.703, is 0.381. Directly above sits MAI-Thinking-1 at 0.350. 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.

What does Claude Haiku 4.5 score on HealthBench Professional?

Claude Haiku 4.5 scores 0.322 on HealthBench Professional, which places it at rank 43 of 43 models on the board as of November 2, 2026. The number was read from Claude Haiku 5.5 System Card, listed on the sources page.

How much does Claude Haiku 4.5 cost to run?

Claude Haiku 4.5 is priced at $1.00 per million input tokens and $5.00 per million output tokens through Anthropic's API.

Pairings with a dedicated comparison page are linked; every other difference is in the score-difference matrix.

  • Claude Haiku 4.5 vs GPT-6 Astra (Anthropic run)
    0.322 vs 0.703, GPT-6 Astra (Anthropic run) by 0.381
  • Claude Haiku 4.5 vs Claude Sonnet 5.5
    0.322 vs 0.692, Claude Sonnet 5.5 by 0.370
  • Claude Haiku 4.5 vs Claude Fable 5
    0.322 vs 0.660, Claude Fable 5 by 0.338
  • Claude Haiku 4.5 vs Claude Opus 5.5
    0.322 vs 0.656, Claude Opus 5.5 by 0.334
  • Claude Haiku 4.5 vs GPT-6 Luna (Anthropic run)
    0.322 vs 0.655, GPT-6 Luna (Anthropic run) by 0.333
  • Claude Haiku 4.5 vs Claude Haiku 5.5
    0.322 vs 0.648, Claude Haiku 5.5 by 0.326
  • Claude Haiku 4.5 vs GPT-6 Astra
    0.322 vs 0.647, GPT-6 Astra by 0.325
  • Claude Haiku 4.5 vs GPT-6.1 Sol
    0.322 vs 0.642, GPT-6.1 Sol by 0.320
  • Claude Haiku 4.5 vs Claude Fable 5 (September card)
    0.322 vs 0.633, Claude Fable 5 (September card) by 0.311
  • Claude Haiku 4.5 vs GPT-6.1 Terra
    0.322 vs 0.627, GPT-6.1 Terra by 0.305
  • Claude Haiku 4.5 vs Muse Wonder Full
    0.322 vs 0.623, Muse Wonder Full by 0.301
  • Claude Haiku 4.5 vs Claude Fable 5.1
    0.322 vs 0.621, Claude Fable 5.1 by 0.299
  • Claude Haiku 4.5 vs GPT-6 Luna
    0.322 vs 0.608, GPT-6 Luna by 0.286
  • Claude Haiku 4.5 vs GPT-6 Sol
    0.322 vs 0.608, GPT-6 Sol by 0.286
  • Claude Haiku 4.5 vs GPT-5.6 Sol
    0.322 vs 0.605, GPT-5.6 Sol by 0.283
  • Claude Haiku 4.5 vs Muse Wonder
    0.322 vs 0.605, Muse Wonder by 0.283
  • Claude Haiku 4.5 vs Claude Opus 5
    0.322 vs 0.598, Claude Opus 5 by 0.276
  • Claude Haiku 4.5 vs Muse Spark 1.1
    0.322 vs 0.593, Muse Spark 1.1 by 0.271
  • Claude Haiku 4.5 vs Claude Sonnet 5
    0.322 vs 0.578, Claude Sonnet 5 by 0.256
  • Claude Haiku 4.5 vs GPT-5.6 Terra
    0.322 vs 0.577, GPT-5.6 Terra by 0.255
  • Claude Haiku 4.5 vs Claude Opus 4.8 (Opus 4.8 grader)
    0.322 vs 0.574, Claude Opus 4.8 (Opus 4.8 grader) by 0.252
  • Claude Haiku 4.5 vs Grok 4.7
    0.322 vs 0.567, Grok 4.7 by 0.245
  • Claude Haiku 4.5 vs Claude Opus 4.8
    0.322 vs 0.558, Claude Opus 4.8 by 0.236
  • Claude Haiku 4.5 vs GPT-5.6 Luna
    0.322 vs 0.557, GPT-5.6 Luna by 0.235
  • Claude Haiku 4.5 vs Gemini 4 Argon
    0.322 vs 0.551, Gemini 4 Argon by 0.229
  • Claude Haiku 4.5 vs Muse Spark
    0.322 vs 0.541, Muse Spark by 0.219
  • Claude Haiku 4.5 vs GPT-5.6 Sol (August)
    0.322 vs 0.540, GPT-5.6 Sol (August) by 0.218
  • Claude Haiku 4.5 vs Kimi K3
    0.322 vs 0.532, Kimi K3 by 0.210
  • Claude Haiku 4.5 vs GLM 5.3
    0.322 vs 0.528, GLM 5.3 by 0.206
  • Claude Haiku 4.5 vs Claude Opus 4.7
    0.322 vs 0.519, Claude Opus 4.7 by 0.197
  • Claude Haiku 4.5 vs GPT-5.5
    0.322 vs 0.518, GPT-5.5 by 0.196
  • Claude Haiku 4.5 vs Gemini 3.8 Flash
    0.322 vs 0.517, Gemini 3.8 Flash by 0.195
  • Claude Haiku 4.5 vs DeepSeek V4.1 Flash
    0.322 vs 0.507, DeepSeek V4.1 Flash by 0.185
  • Claude Haiku 4.5 vs Grok 4.6
    0.322 vs 0.485, Grok 4.6 by 0.163
  • Claude Haiku 4.5 vs GPT-5.4
    0.322 vs 0.481, GPT-5.4 by 0.159
  • Claude Haiku 4.5 vs GPT-5
    0.322 vs 0.462, GPT-5 by 0.140
  • Claude Haiku 4.5 vs GPT-5.2
    0.322 vs 0.459, GPT-5.2 by 0.137
  • Claude Haiku 4.5 vs Claude Sonnet 4.6
    0.322 vs 0.442, Claude Sonnet 4.6 by 0.120
  • Claude Haiku 4.5 vs GPT-5.6 Luna (August)
    0.322 vs 0.441, GPT-5.6 Luna (August) by 0.119
  • Claude Haiku 4.5 vs GPT-5.1
    0.322 vs 0.396, GPT-5.1 by 0.074
  • Claude Haiku 4.5 vs GPT-5.5 Instant
    0.322 vs 0.384, GPT-5.5 Instant by 0.062
  • Claude Haiku 4.5 vs MAI-Thinking-1
    0.322 vs 0.350, MAI-Thinking-1 by 0.028

Where the scores come from and how they are read is on the methodology page. The full ranking is on the leaderboard.