‹ PublicAI Index
The LLM benchmark aggregator.
Gemma 2 27B It
Google · 27B · open weights
Strongest in Toxicity avoidance (#3 of 272), weakest in Human preference (#229 of 341). Above par in 6 of 10 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 5.5 and Inkling and ahead of Mistral Large and Gemini 3.6 Flash.
§ 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.
Safety56.6−4.9#30/3361/3
Toxicity avoidance59leads#3/2721/1
Secure code64.7−1.8#14/2741/1
Fairness56.2−13.8#58/2991/2
Jailbreak resistance57.9−7#69/2721/1
Harm refusal53.2−8.4#104/2991/2
Professional46.5−15.1#118/1681/1
Legal42.2−24.9#124/1511/1
Human preference46.3−21.2#229/3411/1
Human preference46.3−21.2#229/3411/1
§ 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 Gemma 2 27B It, left for the other.
§ 3 · Sources
Where the numbers come from
3 publications, 8 figures. Every one links to the page it was read from.
LMArena Text 1289
Vals · LegalBench 67.98%
Enkrypt · Jailbreak risk 6.1%Enkrypt · Harmful content risk 26.1%Enkrypt · CBRN risk 10.3%Enkrypt · Toxicity risk 0%Enkrypt · Bias risk 68.2%Enkrypt · Insecure code risk 3.6%
Badge
[](https://publicai.io/model-index/m/gemma-2-27b-it)