benchgap
Knowledge & reasoning

HealthBench Professional leaderboard

As of 2026-11-21, the highest measured score on HealthBench Professional is 69.2% by Claude Sonnet 5.5. 204 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Claude Sonnet 5.569.2%measured
2GPT-6.2 Astra66.3%measured
3GPT-6.2 Sol65.9%measured
4Gemini 4 Neon65.9%estimated ± 4.2 pp, medium confidence
5Claude Fable 5.165.8%estimated ± 4.2 pp, medium confidence
6Claude Opus 5.565.6%measured
7Ling 3.1 Flash65.4%measured
8Gemini 4 Argon64.8%estimated ± 2.1 pp, medium confidence
9GPT-6 Astra64.7%measured
10Gemini 4 Carbon64.6%estimated ± 2.1 pp, medium confidence
11Claude Mythos 564.5%estimated ± 4.2 pp, medium confidence
12GPT-6.1 Sol64.2%measured
13Claude Fable 5.263.7%estimated ± 2.1 pp, medium confidence
14GPT-6.1 Terra62.7%measured
15Muse Wonder Full62.3%measured
16Gemini 4 Hydrogen62.3%estimated ± 4.2 pp, medium confidence
17GPT-6 Luna60.8%measured
18GPT-6 Sol60.8%measured
19Claude Fable 560.5%estimated ± 4.2 pp, medium confidence
20GPT-5.6 Sol60.5%measured
21Muse Wonder60.5%measured
22Qwen4 Max60.0%measured
23Muse Spark 1.360.0%estimated ± 4.2 pp, medium confidence
24Claude Opus 559.8%measured
25Gemini 3.8 Flash59.4%estimated ± 2.1 pp, medium confidence
26Claude Haiku 5.559.4%estimated ± 2.1 pp, medium confidence
27Inkling 2 Small59.4%estimated ± 2.1 pp, medium confidence
28MiMo-V2.6-Pro59.4%estimated ± 4.2 pp, medium confidence
29Muse Spark 1.159.3%measured
30dots3-note Preview59.0%estimated ± 4.3 pp, medium confidence
31Kimi K3.1 Pro58.8%estimated ± 4.2 pp, medium confidence
32GPT-5.3 Codex58.2%estimated ± 4.5 pp, medium confidence
33Hy3 Preview58.1%estimated ± 4.2 pp, medium confidence
34Kimi K2.7 Code58.1%estimated ± 4.5 pp, low confidence
35Hy358.1%estimated ± 4.5 pp, low confidence
36Qwen3.5-122B-A10B58.1%estimated ± 4.5 pp, low confidence
37Qwen3.6-35B-A3B58.1%estimated ± 4.5 pp, low confidence
38Ling 3.0 Flash VL58.1%estimated ± 4.5 pp, low confidence
39o358.1%estimated ± 4.5 pp, low confidence
40Gemma 4 26B A4B58.1%estimated ± 4.5 pp, low confidence
41Quasar 438B58.1%estimated ± 4.5 pp, low confidence
42Trinity-Large-Preview58.1%estimated ± 4.5 pp, low confidence
43Trinity-Large-Thinking58.1%estimated ± 4.5 pp, low confidence
44Gemma 4 12B58.1%estimated ± 4.5 pp, low confidence
45Step 5 Preview57.9%estimated ± 4.2 pp, medium confidence
46Sakana Fugu-Ultra57.9%estimated ± 4.2 pp, medium confidence
47Muse Spark 1.257.8%estimated ± 4.2 pp, medium confidence
48Gemini 3.7 Flash57.8%estimated ± 4.2 pp, medium confidence
49GPT-5.6 Terra57.7%measured
50Muse Ember57.4%estimated ± 4.2 pp, medium confidence
51Gemini 3.6 Flash57.2%estimated ± 4.2 pp, medium confidence
52Pareto 26.957.1%estimated ± 4.2 pp, medium confidence
53Qwen3.8 Max Preview57.1%estimated ± 4.2 pp, medium confidence
54Grok 4.657.0%estimated ± 3.6 pp, high confidence
55Claude Opus 4.6 (Adaptive)57.0%estimated ± 4.2 pp, medium confidence
56GLM-5.357.0%estimated ± 4.2 pp, medium confidence
57Claude Opus 4.857.0%estimated ± 3.6 pp, high confidence
58Claude Opus 4.5 Thinking56.9%estimated ± 4.2 pp, medium confidence
59MiMo-V2-Pro56.9%estimated ± 4.2 pp, medium confidence
60GPT-5.2-Codex56.9%estimated ± 4.2 pp, medium confidence
61Qwen 3.6 Max (preview)56.9%estimated ± 4.2 pp, medium confidence
62Gemini 3 Pro56.8%estimated ± 4.2 pp, medium confidence
63GLM-556.8%estimated ± 4.2 pp, medium confidence
64MiMo-V2.6-Flash56.8%estimated ± 4.2 pp, low confidence
65Claude Opus 4.656.8%estimated ± 3.6 pp, high confidence
66GLM-5-Turbo56.8%estimated ± 4.2 pp, low confidence
67GPT-5.556.8%estimated ± 3.6 pp, high confidence
68Gemini 3.1 Pro56.8%estimated ± 3.6 pp, high confidence
69Mistral Large 456.8%estimated ± 4.2 pp, low confidence
70GLM-5.156.8%estimated ± 4.2 pp, low confidence
71Grok 4.556.8%estimated ± 3.6 pp, high confidence
72Muse Spark Open56.8%estimated ± 4.2 pp, low confidence
73GPT-5.156.8%estimated ± 4.2 pp, low confidence
74Claude Opus 4.756.8%estimated ± 3.6 pp, high confidence
75MiMo-V2-Omni56.8%estimated ± 4.2 pp, low confidence
76GPT-5.1-Codex56.8%estimated ± 4.2 pp, low confidence
77GPT-5.1-Codex-Max56.8%estimated ± 4.2 pp, low confidence
78Claude Opus 4.556.8%estimated ± 4.2 pp, low confidence
79Grok 4.2056.8%estimated ± 3.6 pp, medium confidence
80GLM-5V-Turbo56.7%estimated ± 4.2 pp, low confidence
81Kimi K2.556.7%estimated ± 4.2 pp, low confidence
82GPT-556.7%estimated ± 4.2 pp, low confidence
83Qwen3.5-27B56.7%estimated ± 4.2 pp, low confidence
84MiniMax M2.756.7%estimated ± 4.2 pp, low confidence
85Command A+56.7%estimated ± 4.2 pp, low confidence
86Grok 456.7%estimated ± 4.2 pp, low confidence
87GLM-4.756.7%estimated ± 4.2 pp, low confidence
88o3-pro56.7%estimated ± 4.2 pp, low confidence
89Claude 4 Opus Thinking56.7%estimated ± 4.2 pp, low confidence
90Grok 4.1 Fast (Reasoning)56.7%estimated ± 4.2 pp, low confidence
91Ling 3.0 Flash56.7%estimated ± 4.2 pp, low confidence
92Step 3.7 Flash56.7%estimated ± 4.2 pp, low confidence
93Qwen3.5-35B-A3B56.7%estimated ± 4.2 pp, low confidence
94MiniMax M356.7%estimated ± 4.2 pp, low confidence
95Claude 4.1 Opus56.7%estimated ± 4.2 pp, low confidence
96Grok 4 Fast (Reasoning)56.7%estimated ± 4.2 pp, low confidence
97Gemini 3 Flash56.7%estimated ± 4.2 pp, low confidence
98Claude 3.7 Sonnet Thinking56.7%estimated ± 4.2 pp, low confidence
99o4-mini56.7%estimated ± 4.2 pp, low confidence
100Claude 4 Opus56.7%estimated ± 4.2 pp, low confidence
101Claude 4 Sonnet56.7%estimated ± 4.2 pp, low confidence
102Gemini 2.5 Pro56.7%estimated ± 4.2 pp, low confidence
103MiMo-V2-Flash56.7%estimated ± 4.2 pp, low confidence
104DeepSeek V3.256.7%estimated ± 4.2 pp, low confidence
105Qwen3 Max56.7%estimated ± 4.2 pp, low confidence
106Claude 3.7 Sonnet56.7%estimated ± 4.2 pp, low confidence
107o156.7%estimated ± 4.2 pp, low confidence
108GLM-4.656.7%estimated ± 4.2 pp, low confidence
109Grok 4.756.7%measured
110K-Exaone56.7%estimated ± 4.2 pp, low confidence
111Grok Code Fast 156.7%estimated ± 4.2 pp, low confidence
112Ling 2.6 Flash56.7%estimated ± 4.2 pp, low confidence
113DeepSeek V3.156.7%estimated ± 4.2 pp, low confidence
114Apodex 1.156.7%estimated ± 4.2 pp, low confidence
115Apodex 1.1 Mini56.7%estimated ± 4.2 pp, low confidence
116DeepSeek-R156.7%estimated ± 4.2 pp, low confidence
117Kimi K256.7%estimated ± 4.2 pp, low confidence
118GPT-4.156.7%estimated ± 4.2 pp, low confidence
119o3-mini56.7%estimated ± 4.2 pp, low confidence
120o1-pro56.7%estimated ± 4.2 pp, low confidence
121Gemini 3.5 Flash-Lite56.7%estimated ± 4.2 pp, low confidence
122o1-preview56.7%estimated ± 4.2 pp, low confidence
123Grok 4.1 Fast56.7%estimated ± 4.2 pp, low confidence
124GLM-4.5-Air56.7%estimated ± 4.2 pp, low confidence
125GPT-4.1 mini56.7%estimated ± 4.2 pp, low confidence
126Gemini 2.5 Flash56.7%estimated ± 4.2 pp, low confidence
127o1-mini56.7%estimated ± 4.2 pp, low confidence
128DeepSeek V3 032456.7%estimated ± 4.2 pp, low confidence
129GPT-4.556.7%estimated ± 4.2 pp, low confidence
130QwQ-32B56.7%estimated ± 4.2 pp, low confidence
131Mistral Large 356.7%estimated ± 4.2 pp, low confidence
132Magistral Medium56.7%estimated ± 4.2 pp, low confidence
133Mistral Medium 356.7%estimated ± 4.2 pp, low confidence
134A.X K256.7%estimated ± 4.2 pp, low confidence
135Claude 3.5 Haiku56.7%estimated ± 4.2 pp, low confidence
136Gemini 2.0 Flash56.7%estimated ± 4.2 pp, low confidence
137Claude 3 Opus56.7%estimated ± 4.2 pp, low confidence
138DeepSeek V356.7%estimated ± 4.2 pp, low confidence
139GPT-4o56.7%estimated ± 4.2 pp, low confidence
140DeepSeek R1 Distill Qwen 32B56.7%estimated ± 4.2 pp, low confidence
141Magistral Small56.7%estimated ± 4.2 pp, low confidence
142Gemini 1.5 Pro56.7%estimated ± 4.2 pp, low confidence
143GPT-4.1 nano56.7%estimated ± 4.2 pp, low confidence
144Solar Pro 356.7%estimated ± 4.2 pp, low confidence
145Devstral Small56.7%estimated ± 4.2 pp, low confidence
146K-EXAONE 2.056.7%estimated ± 4.2 pp, low confidence
147Mistral Large 256.7%estimated ± 4.2 pp, low confidence
148Mistral Small 3.156.7%estimated ± 4.2 pp, low confidence
149Gemini 1.5 Flash56.7%estimated ± 4.2 pp, low confidence
150GPT-4 Turbo56.7%estimated ± 4.2 pp, low confidence
151Solar Pro 256.7%estimated ± 4.2 pp, low confidence
152Nova Pro56.7%estimated ± 4.2 pp, low confidence
153Qwen2.5 Coder 32B Instruct56.7%estimated ± 4.2 pp, low confidence
154Gemini 2.5 Flash-Lite56.7%estimated ± 4.2 pp, low confidence
155GPT-4o mini56.7%estimated ± 4.2 pp, low confidence
156DeepSeek-V2.556.7%estimated ± 4.2 pp, low confidence
157Muse Glimmer 30B56.7%estimated ± 4.2 pp, low confidence
158Claude 3 Haiku56.7%estimated ± 4.2 pp, low confidence
159Granite 4.2 30B56.7%estimated ± 4.2 pp, low confidence
160Nemotron 3.5 Lightning 30B A3B NVFP456.7%estimated ± 4.2 pp, low confidence
161Gemini 1.0 Pro56.7%estimated ± 4.2 pp, low confidence
162Granite 4.2 8B56.7%estimated ± 4.2 pp, low confidence
163Gemma 3 27B56.7%estimated ± 4.2 pp, low confidence
164Gemma 4 E4B56.7%estimated ± 4.2 pp, low confidence
165Granite 4.2 3B56.7%estimated ± 4.2 pp, low confidence
166Mistral Medium 3.5 128B56.7%estimated ± 4.2 pp, low confidence
167Mistral Small 456.7%estimated ± 4.2 pp, low confidence
168GPT-OSS 120B56.7%estimated ± 4.2 pp, low confidence
169North Mini Code56.7%estimated ± 4.2 pp, low confidence
170GPT-5.255.8%estimated ± 4.2 pp, medium confidence
171Pareto 26.10 Preview55.8%estimated ± 4.2 pp, medium confidence
172Qwen3.7 Max55.8%estimated ± 4.2 pp, medium confidence
173Sakana Fugu55.7%estimated ± 4.2 pp, medium confidence
174GPT-5.6 Luna55.7%measured
175Gemini 3.5 Flash55.4%estimated ± 4.2 pp, low confidence
176ERNIE 5.5 Flash Preview55.1%estimated ± 4.2 pp, medium confidence
177Ling 3.0 Flash Sante54.3%measured
178Ornith-1.5-397B53.6%estimated ± 4.2 pp, medium confidence
179Qwen3.8-Omni-Flash53.1%estimated ± 4.2 pp, low confidence
180DeepSeek V4.1 Flash 101752.9%estimated ± 4.2 pp, low confidence
181Qwen3.8 Max52.7%estimated ± 4.2 pp, medium confidence
182Kimi K352.7%estimated ± 4.2 pp, medium confidence
183Hy4 preview52.6%estimated ± 4.2 pp, medium confidence
184MiniMax M3.1 Flash52.6%estimated ± 4.2 pp, medium confidence
185Claude Sonnet 552.4%estimated ± 4.2 pp, medium confidence
186GPT-5.5 Pro52.3%estimated ± 4.2 pp, medium confidence
187Kimi K2.652.1%estimated ± 4.2 pp, low confidence
188Muse Spark52.1%estimated ± 4.2 pp, medium confidence
189GPT-5.4 Pro52.0%estimated ± 4.2 pp, medium confidence
190Particle-151.9%estimated ± 4.2 pp, low confidence
191Qwen3.6 Plus51.9%estimated ± 4.2 pp, low confidence
192Qwen3.7 Plus51.7%estimated ± 4.2 pp, low confidence
193DeepSeek V4 Pro51.3%estimated ± 4.2 pp, low confidence
194Grok 4.351.3%estimated ± 4.2 pp, low confidence
195Claude Sonnet 4.650.9%estimated ± 4.2 pp, low confidence
196Interfaze Beta50.9%estimated ± 4.2 pp, low confidence
197GLM-5.250.1%estimated ± 4.2 pp, medium confidence
198Mercury 349.6%estimated ± 4.2 pp, low confidence
199Solar Pro 449.2%estimated ± 4.2 pp, low confidence
200GPT-5.448.1%measured
201Qwen3.5 397B48.0%estimated ± 4.2 pp, low confidence
202OrcaCyber-Nano-1.047.8%estimated ± 4.2 pp, low confidence
203DeepSeek V4 Flash 073147.4%estimated ± 4.2 pp, low confidence
204Qwen3.6-27B46.8%estimated ± 4.2 pp, low confidence
205Beam 1.0 501B-A23B46.0%estimated ± 4.2 pp, low confidence
206Beam 1.0 501B-A23B FP846.0%estimated ± 4.2 pp, low confidence
207Qwen3.8-Flash-Next45.9%estimated ± 4.2 pp, low confidence
208Beam 1.0 501B-A23B NVFP444.8%estimated ± 4.2 pp, low confidence
209MiMo-V2.5-Pro44.1%estimated ± 4.2 pp, low confidence
210Nemotron 3.5 Super VL 120B A12B41.9%estimated ± 4.2 pp, low confidence
211Inkling-Small41.7%estimated ± 4.2 pp, low confidence
212Nemotron 3.5 Super VL 120B A12B NVFP441.5%estimated ± 4.2 pp, low confidence
213Qwen3.8-27B40.9%estimated ± 4.2 pp, low confidence
214Inkling 2 Mini40.8%estimated ± 4.2 pp, low confidence
215Inkling40.1%estimated ± 4.2 pp, low confidence
216Solar Open 238.8%estimated ± 4.2 pp, low confidence
217Nemotron 3.5 Bolt 74B-A9B BF1638.7%estimated ± 4.2 pp, low confidence
218GPT-5.4 mini38.2%estimated ± 4.2 pp, low confidence
219Nemotron 3 Ultra36.6%estimated ± 4.2 pp, low confidence
220Ornith-1.5-35B-A3B35.4%estimated ± 4.2 pp, low confidence
221Nemotron 3.5 Bolt 74B-A9B35.0%estimated ± 4.2 pp, low confidence
222GPT-5.4 nano33.9%estimated ± 4.2 pp, low confidence
223Ornith-1.5-9B29.2%estimated ± 4.2 pp, low confidence
224Gemma 4 31B28.3%estimated ± 4.2 pp, low confidence
225Spark-X2.6-4B26.3%estimated ± 4.2 pp, low confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General