‹ PublicAI Index
The LLM benchmark aggregator.
GPT-4o
OpenAI
Strongest in Multimodal understanding (#1 of 19), weakest in Agents (#247 of 298). Above par in 14 of 26 scopes. Among the models it meets almost everywhere, it finishes behind Grok 4.7 and Claude Fable 5 and ahead of Grok 4.1 Fast and Claude 3.5 Haiku.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Safety55.2−6.2#48/3293/3
Safe-prompt compliance57.1−3.3#20/821/1
Toxicity avoidance56.7−2.3#50/2681/1
Factual grounding50.9−19.8#54/1041/1
Secure code61−5.2#55/2701/1
Jailbreak resistance59.1−5.9#56/2681/1
Harm refusal54.6−6.9#76/2922/2
Fairness52.7−17.3#88/2922/2
Knowledge53.5−9#51/1972/3
Multimodal understanding72.1leads#1/191/1
Academic knowledge39.2−24.7#104/1231/1
Professional49.5−12.9#104/1941/2
Medical52.5−14#71/1491/1
Legal52.5−13.4#73/1731/1
Finance45.9−16.5#120/1621/1
Coding40.6−25.6#163/1792/6
Code generation33.4−31.9#16/201/1
Agentic coding44−22.9#127/1731/5
Reasoning34.9−30.5#190/1953/5
Science34.5−29#105/1221/1
Mathematics35.5−28.3#138/1521/2
Reasoning34.7−34.3#144/1472/3
Human preference50.5−18.3#191/3541/1
Human preference50.5−18.3#191/3541/1
Agents43.1−23.3#247/2981/6
Knowledge work36.1−34.6#174/2081/2
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for GPT-4o, left for the other.
§ 3 · Sources
Where the numbers come from
11 publications, 28 figures. Every one links to the page it was read from.
LMArena Text 1336
GDPval-AA -9
ARC-AGI-2 0%
LiveCodeBench 29.5%
Aider polyglot 18.2%
MMMU-Pro 51.9%
SimpleBench 17.8%
Vals · CaseLaw 59.7%Vals · LegalBench 82.21%Vals · CorpFin 45.92%Vals · TaxEval 74.53%Vals · MortgageTax 57.43%Vals · MedQA 88.16%Vals · GPQA Diamond 53.79%Vals · MMLU Pro 72.56%Vals · AIME 11.88%
HELM Safety · HarmBench 82.9%HELM Safety · SimpleSafetyTests 98.5%HELM Safety · Anthropic Red Team 99.1%HELM Safety · BBQ 95.1%HELM Safety · XSTest 97.3%
Enkrypt · Jailbreak risk 5.2%Enkrypt · Harmful content risk 32.2%Enkrypt · CBRN risk 6.2%Enkrypt · Toxicity risk 1.6%Enkrypt · Bias risk 79.3%Enkrypt · Insecure code risk 10.7%
Vectara · Factual consistency 90.4%
Badge
[](https://publicai.io/model-index/m/gpt-4o)