‹ PublicAI Index
The LLM benchmark aggregator.
Qwen3.5 35B A3B
Alibaba · 36B · A3B · open weights
Strongest in Secure code (#40 of 268), weakest in Jailbreak resistance (#176 of 266). Above par in 7 of 11 scopes. Among the models it meets almost everywhere, it finishes behind Grok 4.7 and Claude Opus 5.5 and ahead of Kimi K2.6 and Kimi K2.5.
§ 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.
Safety53.2−8.2#93/3272/3
Secure code62.3−3.9#40/2681/1
Toxicity avoidance56.6−2.4#50/2661/1
Factual grounding48.6−22.1#62/1041/1
Fairness52.8−17.2#83/2901/2
Harm refusal52−9.5#123/2901/2
Jailbreak resistance49.3−15.7#176/2661/1
Human preference56.1−12.7#139/3541/1
Human preference56.1−12.7#139/3541/1
Agents48.2−18.1#162/2981/6
Knowledge work46.4−24.3#119/2081/2
§ 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 Qwen3.5 35B A3B, left for the other.
§ 3 · Sources
Where the numbers come from
4 publications, 9 figures. Every one links to the page it was read from.
LMArena Text 1394
GDPval-AA 616
Enkrypt · Jailbreak risk 13.4%Enkrypt · Harmful content risk 1.7%Enkrypt · CBRN risk 26.8%Enkrypt · Toxicity risk 1.6%Enkrypt · Bias risk 73.4%Enkrypt · Insecure code risk 8%
Vectara · Factual consistency 89.5%
Badge
[](https://publicai.io/model-index/m/qwen3-5-35b-a3b)