‹ PublicAI Index
The LLM benchmark aggregator.
GPT-4 Turbo
OpenAI
Strongest in Safe-prompt compliance (#15 of 82), weakest in Human preference (#203 of 354). Above par in 9 of 16 scopes. Among the models it meets almost everywhere, it finishes behind Claude Sonnet 5 and Claude Opus 4.8 and ahead of GPT-4.1 mini and GPT OSS 120B.
§ 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.
Safety57.1−4.3#24/3292/3
Safe-prompt compliance57.9−2.5#15/821/1
Harm refusal56.9−4.6#33/2922/2
Jailbreak resistance60.6−4.4#34/2681/1
Toxicity avoidance56.5−2.5#56/2681/1
Secure code60.8−5.4#57/2701/1
Fairness54.4−15.6#72/2922/2
Professional50.3−12.1#91/1941/2
Legal51.7−14.2#78/1731/1
Medical49.2−17.3#94/1491/1
Coding42.6−23.6#147/1791/6
Code generation32.9−32.4#17/201/1
Reasoning42.8−22.6#155/1951/5
Reasoning38.4−30.6#126/1471/3
Human preference49.5−19.3#203/3541/1
Human preference49.5−19.3#203/3541/1
§ 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-4 Turbo, left for the other.
§ 3 · Sources
Where the numbers come from
6 publications, 16 figures. Every one links to the page it was read from.
LMArena Text 1325
LiveCodeBench 28.7%
SimpleBench 25.1%
Vals · LegalBench 80.46%Vals · MedQA 81.99%
HELM Safety · HarmBench 89.8%HELM Safety · SimpleSafetyTests 99%HELM Safety · Anthropic Red Team 99.7%HELM Safety · BBQ 94.1%HELM Safety · XSTest 97.7%
Enkrypt · Jailbreak risk 3.9%Enkrypt · Harmful content risk 23.9%Enkrypt · CBRN risk 5.3%Enkrypt · Toxicity risk 1.7%Enkrypt · Bias risk 73.6%Enkrypt · Insecure code risk 11.1%
Badge
[](https://publicai.io/model-index/m/gpt-4-turbo)