‹ PublicAI Index
The LLM benchmark aggregator.
GPT-4.1 mini
OpenAI
Strongest in Safe-prompt compliance (#18 of 82), weakest in Toxicity avoidance (#176 of 268). Above par in 12 of 26 scopes. Among the models it meets almost everywhere, it finishes behind GPT-5.2 and Claude Opus 5.5 and ahead of MiMo V2.5 Pro and Gemini 2.5 Flash-Lite.
§ 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.
Professional51.2−11.2#84/1941/2
Finance52.4−10#81/1621/1
Medical50.6−15.9#82/1491/1
Legal49.8−16.1#90/1731/1
Safety53.2−8.2#91/3292/3
Safe-prompt compliance57.3−3.1#18/821/1
Secure code59−7.2#71/2701/1
Harm refusal53.5−8#97/2922/2
Fairness49.1−20.9#125/2922/2
Jailbreak resistance52.4−12.6#151/2681/1
Toxicity avoidance52−7#176/2681/1
Coding47.6−18.6#104/1791/6
Agentic coding47.3−19.6#102/1731/5
Core abilities48.6−16.7#115/1991/4
General intelligence48−21.4#109/1781/3
Knowledge47.9−14.6#136/1971/3
Academic knowledge45−18.9#97/1231/1
Agents49.4−17#137/2982/6
Tool use59.1−14.9#22/811/1
Knowledge work40.8−29.9#147/2081/2
Reasoning44.1−21.3#148/1952/5
Science44.3−19.2#93/1221/1
Mathematics45.5−18.3#105/1521/2
Reasoning41.9−27.1#113/1471/3
Human preference55−13.8#152/3541/1
Human preference55−13.8#152/3541/1
§ 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-4.1 mini, left for the other.
§ 3 · Sources
Where the numbers come from
9 publications, 25 figures. Every one links to the page it was read from.
LMArena Text 1383
GDPval-AA 277
ARC-AGI-2 0%
Aider polyglot 32.4%
BFCL v4 50.45%
Kagi LLM Benchmark 48.6%
Vals · LegalBench 78.04%Vals · CorpFin 57.93%Vals · TaxEval 71.91%Vals · MortgageTax 65.5%Vals · MedQA 84.63%Vals · GPQA Diamond 67.93%Vals · MMLU Pro 77.22%Vals · AIME 49.38%
HELM Safety · HarmBench 85.6%HELM Safety · SimpleSafetyTests 100%HELM Safety · Anthropic Red Team 99.3%HELM Safety · BBQ 92.1%HELM Safety · XSTest 97.4%
Enkrypt · Jailbreak risk 10.8%Enkrypt · Harmful content risk 41.7%Enkrypt · CBRN risk 11.3%Enkrypt · Toxicity risk 4.8%Enkrypt · Bias risk 84.5%Enkrypt · Insecure code risk 14.7%
Badge
[](https://publicai.io/model-index/m/gpt-4-1-mini)