Benchmark Almanac

OCT-Bench

GPT-5.4 Mini 62% leads since July 2026
Not in model cards

Models answer 10,076 questions about retinal scans, from anatomy to treatment and prognosis. The evaluator measures answer accuracy and counts invalid responses as incorrect.

Questions cover 20 tasks over 4,137 optical coherence tomography images. The scope includes clinical reasoning beyond image classification.

Released
July 2026
Kind
Medicine
Level
Expert
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Images → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-5.4 Mini #74 on the capability index
Last model added
March 2026
Scores refreshed
November 18th, 2026

In the authors’ words

Optical coherence tomography (OCT) imaging is essential for the diagnosis and treatment of retinal diseases. Although multimodal large language models (MLLMs) have demonstrated considerable potential in medical image analysis, existing benchmarks largely reduce OCT understanding to coarse-grained disease classification or isolated visual question answering, leaving the complete cognitive process from visual perception to clinical reasoning insufficiently evaluated. To address this limitation, we introduce OCT-Bench, a comprehensive benchmark dedicated to OCT image understanding. OCT-Bench comprises 10,076 high-quality multiple-choice questions constructed from 4,137 OCT images across seven public datasets. Following the real-world clinical interpretation workflow, we establish a hierarchical capability taxonomy consisting of 20 fine-grained tasks across three dimensions: Perception, Cognition, and Reasoning. These tasks cover a broad range of capabilities, including imaging attributes, retinal anatomy, lesion characteristics, spatial relationships, disease assessment, therapeutic decision-making, and prognostic management. We systematically evaluate 20 representative MLLMs, including proprietary models, open-source general-purpose models, and medical-domain models. Experimental results demonstrate that current models remain substantially short of reliable OCT understanding. Moreover, neither medical-domain adaptation nor increased model scale consistently improves performance across capability levels. OCT-Bench enables comprehensive and fine-grained evaluation of MLLMs, providing a foundation for identifying capability bottlenecks and advancing clinically grounded OCT understanding.

Abstract of Can Multimodal Large Language Models Understand OCT? (arXiv:2607.16609, CC0)
20
Entries in its paper, 9 models we track

GPT-5.4 Mini leads with 62% among models released by March 2026 GPT-5.4 Mini leads with 62% among models released by March 2026

Best score per entry from the benchmark's paper, read on November 18th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in March 2026, 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-5.4 Mini 62%
  2. 2 Gemini 2.5 Flash Sep 2025 60.5%
  3. 3 Hulu-Med-32B 58.4%
  4. 4 Lingshu-32B 56.6%
  5. 5 HealthGPT-Pro-8B 55.6%
  6. 6 InternVL2.5-26B 54.5%
  7. 7 Hulu-Med-7B 54.1%
  8. 8 Fleming-VL-8B 53.6%
  9. 9 Qwen2.5-VL-32B 52.4%
  10. 10 Grok 4 Fast 52.4%
  11. 11 MediX-R1-8B 51.7%
  12. 12 Lingshu-7B 51.5%
  13. 13 LLaVA-1.6-34B 50.5%
  14. 14 Qwen2.5-VL-7B 49.5%
  15. 15 InternVL2.5-4B 46.6%
  16. 16 HealthGPT-M3 45.1%
  17. 17 Gemma 3 12B 44.7%
  18. 18 MedGemma-4B 42.2%
  19. 19 Phi-3-Vision-128k 42%
  20. 20 LLaVA-v1.5-13B 39.6%

Source: The OCT-Bench paper

#EntryScoreSettingDateSource
1 GPT-5.4 Mini 62% Mar 2026 Official leaderboard
2 Gemini 2.5 Flash Sep 2025 60.5% Sep 2025 Official leaderboard
3 Hulu-Med-32B 58.4% Official leaderboard
4 Lingshu-32B 56.6% Official leaderboard
5 HealthGPT-Pro-8B 55.6% Official leaderboard
6 InternVL2.5-26B 54.5% Dec 2024 Official leaderboard
7 Hulu-Med-7B 54.1% Official leaderboard
8 Fleming-VL-8B 53.6% Official leaderboard
9 Qwen2.5-VL-32B 52.4% Mar 2025 Official leaderboard
10 Grok 4 Fast 52.4% Sep 2025 Official leaderboard
11 MediX-R1-8B 51.7% Official leaderboard
12 Lingshu-7B 51.5% Official leaderboard
13 LLaVA-1.6-34B 50.5% Official leaderboard
14 Qwen2.5-VL-7B 49.5% Jan 2025 Official leaderboard
15 InternVL2.5-4B 46.6% Dec 2024 Official leaderboard
16 HealthGPT-M3 45.1% Official leaderboard
17 Gemma 3 12B 44.7% Mar 2025 Official leaderboard
18 MedGemma-4B 42.2% Official leaderboard
19 Phi-3-Vision-128k 42% Official leaderboard
20 LLaVA-v1.5-13B 39.6% Oct 2023 Official leaderboard
#ModelScoreSettingDateSource
1 GPT-5.4 Mini 62% Mar 2026 Official leaderboard
2 Gemini 2.5 Flash Sep 2025 60.5% Sep 2025 Official leaderboard
3 InternVL2.5-26B 54.5% Dec 2024 Official leaderboard
4 Qwen2.5-VL-32B 52.4% Mar 2025 Official leaderboard
5 Grok 4 Fast 52.4% Sep 2025 Official leaderboard
6 Qwen2.5-VL-7B 49.5% Jan 2025 Official leaderboard
7 InternVL2.5-4B 46.6% Dec 2024 Official leaderboard
8 Gemma 3 12B 44.7% Mar 2025 Official leaderboard
9 LLaVA-v1.5-13B 39.6% Oct 2023 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s paper: Perception, Cognition and Reasoning. Each ranks models on one part of the test. GPT-5.4 Mini leads 2 of them, and Gemini 2.5 Flash leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Perception

9 models we track have a score here.

  1. 1 GPT-5.4 Mini 75.8%
  2. 2 InternVL2.5-26B ↑1 72.5%
  3. 3 Gemini 2.5 Flash Sep 2025 ↓1 67.7%
  4. 4 Qwen2.5-VL-7B ↑2 66.7%
  5. 5 Qwen2.5-VL-32B ↓1 66%
  6. 6 Grok 4 Fast ↓1 63.5%
  7. 7 Gemma 3 12B ↑1 62.6%
  8. 8 InternVL2.5-4B ↓1 61.2%
  9. 9 LLaVA-v1.5-13B 54.2%

Number one over time

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

OpenAI
Benchmark paper Not tested since Jul 2026

Source: The OCT-Bench 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 2025 2026 Near the ceiling (85%) GPT-5.4 Mini Gemini 2.5 Flash InternVL2.5-26B LLaVA-v1.5-13B 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) GPT-5.4 Mini InternVL2.5-26B

Source: Official OCT-Bench 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.

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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 8 months ago, in March 2026, for GPT-5.4 Mini. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (20 entries).
  • Pass Models scored 9 models we track, plus 11 other entries on its leaderboard.
  • Warning Newest score The newest came 8 months ago, in March 2026, for GPT-5.4 Mini.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.4 Mini (#74).
  • Pass Headroom The best is 62%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 9 rows for models we track and 11 other entries at the last read.
  • Pass Last read November 18th, 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