Medical AI Superintelligence Test (MAST) Leaderboard

ARISE

The MAST project seeks to curate a centralized resource of robust and realistic clinical benchmarks to measure the performance of medical AI.

See our methodology and submission instructions.

Dec 2025
First Do NOHARM v1SCT-Bench
Feb 2026
MedAgentBench v2
Mar 2026
ReXrank MiniFirst Do NOHARM v2
Apr 2026
CPC-BenchMultimodal Derm
May 2026
PhysicianBench
~Jul 2026
In Development
NOHARM-Mind
~H2 2026 – 2027
In Development
PACT: 12 high-risk clinical reasoning benchmarks

Benchmark Demos

View all ›
First Do NOHARM v2
SCT-Bench
MedAgentBench v2
PhysicianBench
ReXrank Mini
CPC-Bench
CPC-Bench
First Do NOHARM v2
SCT-Bench
MedAgentBench v2
PhysicianBench
ReXrank Mini
CPC-Bench
CPC-Bench
First Do NOHARM v2
SCT-Bench
MedAgentBench v2
PhysicianBench
ReXrank Mini
CPC-Bench
CPC-Bench
First Do NOHARM v2
SCT-Bench
MedAgentBench v2
PhysicianBench
ReXrank Mini
CPC-Bench
CPC-Bench
Preview: MAST is currently in preview. Exact scores on this benchmark may change as we undergo final validation and tuning.

First Do NOHARM v2 overall metric across 45 models

Benchmark Comparison

Overall scores across benchmarks

R² = 0.25

Model Profiles

Compare models across benchmarks

Not all models could be run on every benchmark; axes with no result are shown at the center.

Performance Over Time

No scored data for the current selection.
No scored data for the current selection.
Preview: MAST is currently in preview. Exact scores on this benchmark may change as we undergo final validation and tuning.

Model Leaderboard

Top 10 shown

#ModelReasoning↓SafetyAgenticImagesMultimodalCPCDiagnosticManagementRadiology
1
GPT-5.5
75.4%±2.3
73.4%±4.6
56.7%±4.6
42.9%±2.8
49.2%±1.8
86.2%±3.8
75.1%±3.2
76.2%±2.6
57.6%±0.0
2
GPT-5
72.1%±2.6
67.2%±5.1
32.4%±10.9
44.2%±2.5
47.6%±1.5
80.3%±4.7
73.1%±4.1
73.2%±2.8
51.6%±0.0
3
GPT-5.2
71.8%±2.3
70.6%±4.7
27.2%±10.3
46.5%±2.7
49.3%±1.5
82.8%±3.9
72.6%±3.4
71.7%±2.9
52.4%±0.0
4
GPT-5.4
71.3%±2.2
72.4%±4.4
38.8%±5.2
46.5%±2.5
47.1%±1.3
82.5%±4.0
71.9%±3.4
70.6%±2.8
47.8%±0.0
5
Claude Opus 4.7
70.1%±2.7
67.2%±5.0
40.6%±5.0
42.7%±2.3
48.0%±1.5
76.7%±4.9
75.4%±3.6
67.9%±3.5
54.7%±0.0
6
Kimi K2.6OSS
66.3%±2.9
63.9%±5.0
------
73.9%±5.5
73.1%±4.0
63.1%±3.8
--
7
Claude Opus 4.6
66.1%±3.0
57.9%±5.4
44.4%±5.4
40.4%±2.7
42.7%±1.5
76.0%±5.0
75.4%±3.7
63.8%±3.7
45.2%±0.0
8
GPT-5 mini
66.0%±2.4
62.1%±4.9
19.0%±10.7
43.2%±2.4
46.1%±1.4
77.9%±4.7
68.2%±3.8
65.9%±2.9
49.4%±0.0
9
Claude Sonnet 4.6
64.8%±2.9
57.7%±5.4
35.0%±5.7
39.0%±2.8
39.7%±1.5
76.2%±4.8
72.7%±3.5
62.9%±3.5
40.4%±0.0
10
Kimi K2.5OSS
63.4%±2.9
60.7%±5.0
11.1%±8.3
41.7%±2.8
47.0%±1.8
68.7%±5.5
69.3%±4.3
60.8%±4.0
53.8%±0.0

Not all models could be run on every benchmark; blank (NA) cells indicate no result, not a zero score.