Benchmark Almanac

MedCalc-Bench (original)

GPT-4 (snapshot unspecified) 50.9% tops its leaderboard Llama 3-70B 39.5% leads the models we track
Not in model cards

Models apply 55 clinical calculators to patient notes. The original test contains 1,100 questions and checks numerical answers against task-specific tolerance bounds.

In June 2024, GPT-4 led the paper’s one-shot prompting comparison at 50.91% accuracy. The Verified release later corrected reference labels.

Released
June 2024
Kind
Medicine
Level
Expert
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Items from
Real work and use
Run by
Anyone
Strongest model tested
Llama 3-70B #182 on the capability index
Last model added
April 2024
Scores refreshed
November 20th, 2026
8
Entries in its paper, 4 models we track

GPT-4 (snapshot unspecified) tops the leaderboard with 50.9% Llama 3-70B leads with 39.5% among models released by April 2024

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Llama 3-70B, at 39.5%. This view shows only models we track, each at its best.

The newest model with a score here came out in April 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-4 (snapshot unspecified) One-shot CoT Prompting 50.9%
  2. 2 Llama 3-70B One-shot CoT Prompting 39.5%
  3. 3 GPT-3.5 (snapshot unspecified) One-shot CoT Prompting 34.9%
  4. 4 Mixtral 8x7B One-shot CoT Prompting 29.2%
  5. 5 Llama 3-8B One-shot CoT Prompting 27.1%
  6. 6 MediTron 70B One-shot CoT Prompting 26.3%
  7. 7 Mistral-7B-Instruct-v0.2 One-shot CoT Prompting 16.1%
  8. 8 PMC-LLaMA 13B One-shot CoT Prompting 6.97%

Source: The MedCalc-Bench (original) paper

#EntryScoreSettingDateSource
1 GPT-4 (snapshot unspecified) 50.9% One-shot CoT Prompting Official leaderboard
2 Llama 3-70B 39.5% One-shot CoT Prompting Apr 2024 Official leaderboard
3 GPT-3.5 (snapshot unspecified) 34.9% One-shot CoT Prompting Official leaderboard
4 Mixtral 8x7B 29.2% One-shot CoT Prompting Dec 2023 Official leaderboard
5 Llama 3-8B 27.1% One-shot CoT Prompting Apr 2024 Official leaderboard
6 MediTron 70B 26.3% One-shot CoT Prompting Official leaderboard
7 Mistral-7B-Instruct-v0.2 16.1% One-shot CoT Prompting Dec 2023 Official leaderboard
8 PMC-LLaMA 13B 6.97% One-shot CoT Prompting Official leaderboard
#ModelScoreSettingDateSource
1 Llama 3-70B 39.5% One-shot CoT Prompting Apr 2024 Official leaderboard
2 Mixtral 8x7B 29.2% One-shot CoT Prompting Dec 2023 Official leaderboard
3 Llama 3-8B 27.1% One-shot CoT Prompting Apr 2024 Official leaderboard
4 Mistral-7B-Instruct-v0.2 16.1% One-shot CoT Prompting Dec 2023 Official leaderboard

Rankings by part

We read 7 parts from the benchmark’s paper, each ranking models on one part of the test, such as Lab, Physical and Date. Llama 3-70B leads 6 of them, and Mixtral 8x7B leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Lab

4 models we track have a score here.

  1. 1 Llama 3-70B 41.6%
  2. 2 Llama 3-8B ↑1 34.9%
  3. 3 Mixtral 8x7B ↓1 28.1%
  4. 4 Mistral-7B-Instruct-v0.2 11%

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Llama 3-70B
Zero-shot D… Zero-shot C… One-shot Co…
39.5%
Mixtral 8x7B
Zero-shot D… Zero-shot C… One-shot Co…
29.2%
Llama 3-8B
Zero-shot D… Zero-shot C… One-shot Co…
27.1%
Mistral-7B-Instruct-v0.2
Zero-shot D… One-shot Co…
16.1%

Source: The MedCalc-Bench (original) paper

ModelEffortHarnessScoreCost per taskOutput tokens
Llama 3-70B One-shot CoT Prompting 39.5%
Llama 3-70B Zero-shot CoT Prompting 35.5%
Mixtral 8x7B One-shot CoT Prompting 29.2%
Llama 3-8B One-shot CoT Prompting 27.1%
Mixtral 8x7B Zero-shot CoT Prompting 22.4%
Llama 3-70B Zero-shot Direct Prompting 20.8%
Llama 3-8B Zero-shot CoT Prompting 16.4%
Mistral-7B-Instruct-v0.2 One-shot CoT Prompting 16.1%
Mixtral 8x7B Zero-shot Direct Prompting 14.2%
Llama 3-8B Zero-shot Direct Prompting 13.1%
Mistral-7B-Instruct-v0.2 Zero-shot CoT Prompting 10.8%
Mistral-7B-Instruct-v0.2 Zero-shot Direct Prompting 9.84%

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Llama 3-70B
Zero-shot D… Zero-shot C… One-shot Co…
39.5%
Mixtral 8x7B
Zero-shot D… Zero-shot C… One-shot Co…
29.2%
Llama 3-8B
Zero-shot D… Zero-shot C… One-shot Co…
27.1%
Mistral-7B-Instruct-v0.2
Zero-shot D… One-shot Co…
16.1%

Source: The MedCalc-Bench (original) paper

ModelEffortHarnessScoreCost per taskOutput tokens
Llama 3-70B One-shot CoT Prompting 39.5%
Llama 3-70B Zero-shot CoT Prompting 35.5%
Mixtral 8x7B One-shot CoT Prompting 29.2%
Llama 3-8B One-shot CoT Prompting 27.1%
Mixtral 8x7B Zero-shot CoT Prompting 22.4%
Llama 3-70B Zero-shot Direct Prompting 20.8%
Llama 3-8B Zero-shot CoT Prompting 16.4%
Mistral-7B-Instruct-v0.2 One-shot CoT Prompting 16.1%
Mixtral 8x7B Zero-shot Direct Prompting 14.2%
Llama 3-8B Zero-shot Direct Prompting 13.1%
Mistral-7B-Instruct-v0.2 Zero-shot CoT Prompting 10.8%
Mistral-7B-Instruct-v0.2 Zero-shot Direct Prompting 9.84%

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Meta
Benchmark paper Not tested since Jun 2024
Llama 3-70B 39.5%

Source: The MedCalc-Bench (original) paper

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) Llama 3-70B Mixtral 8x7B 0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) Llama 3-70B Mixtral 8x7B

Source: Official MedCalc-Bench (original) leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

Versions

More in Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the benchmark’s paper (8 entries).
  • Pass Models scored 4 models we track, plus 4 other entries on its leaderboard.
  • Note Newest score The newest came 2 years ago, in April 2024, for Llama 3-70B. That’s expected for a test replaced by MedCalc-Bench Verified.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Llama 3-70B (#182).
  • Pass Headroom The best is 39.5%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 12 rows for models we track and 4 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check